[fitswcs] Re: [iaufwg] Thoughts about time coordinates

[William Thompson]

Steve Allen wrote:
> On Mon 2005-07-11T14:28:05 -0400, William Thompson hath writ:
> 
>>              TDB     Barycentric Dynamical Time (TDT + periodic (~ 10
>>              msec))
> 
> 
> TDB doesn't really exist.  In practice it is always a version of T_{eph}.
> http://www.ucolick.org/~sla/leapsecs/timescales.html#Teph
> That makes it necessary to qualify what ephemeris was implied.
> 
> There is also the upcoming matter of refining the meaning of TT when
> time transfer becomes precise enough that the diurnal GR variation in
> clock rate becomes visible.  At that point it will no longer be
> possible to pretend that a single TT or TAI can describe all points on
> earth.

I'm sure you're right.  The ephemeris program that I'm familiar with is JPL's 
SPICE software, whose documentation tends to use the terms ET and TDB 
interchangeably.

> Aside from System, Format, and Usage, there is the matter of reference
> frame.  The spectral WCS Paper III begins to touch on this,  but it
> explicitly disavows spacecraft.

Good point.

>>      Many types of data analysis depend more on relative time than on
>>      absolute time
> 
> 
> Many types of observation systems have never had requirements for
> accurate timing of shutter operations.  On a good day our CCD systems
> at Lick and Keck are good to a second, on a bad day the system loading
> can make them as bad as 30 seconds.  Nobody demanded anything better
> at the time of the design.

Evidently, that's an example of where one type of astronomy has different 
requirements from another type of astronomy.  I want to make sure that the needs 
of both, where relevant, are met.  In solar astronomy, relative time is used to 
study wave phenomenon.  For example, you might position the aperture of a slit 
spectrograph over active region loops, and measure velocity shifts every few 
seconds over a period of several hours, looking for wave signatures.  You don't 
need to know the absolute time to very high precision, but you do need to know 
the relative timing between the exposures.

>>              TIME can be in any of the time units in Paper I.
> 
> 
> Paper I does not distinguish SI second from mean solar seconds, nor
> does it consider any of the implications of the changes in length of
> TAI seconds, nor does it begin to consider reference frame issues.

No, but each type of time has the concept of a second embedded within it.  All I 
was trying to say was that TIME could be in units of seconds, or minutes, or 
hours, etc., so long as the CUNITia keyword was properly formatted.  The concept 
of a second (or minute, ...) would have to be refined by additional keywords 
(e.g. TIMESYS).

There's also the concept of observatory time (e.g. spacecraft clock time), which 
is generally an approximation to one of the standard definitions of time.  Note 
that this also gets into the distinction between absolute and relative time, 
since it tends to have more of an effect on the former.

>>              In other words, with the UTC projection, every day is treated
>>              as being exactly 86400 seconds long.
> 
> 
> Which means they are effectively mean solar seconds from the inception
> of UTC until now.  But if the Time Lords succeed in redefining UTC
> then they will become TAI seconds at some point in the future.  In a
> historical sense UTC would become even more schizophrenic than it is
> now.

Since 1972, the length of the UTC second and the TAI second have been the same, 
with an occasional leap second added to the UTC time scale.  My understanding of 
the current proposal is that leap seconds would no longer be added after a 
certain date, breaking the rough relationship between UTC and UT1, but 
maintaining the precise relationship between UTC and TAI.

> There is a somewhat serious proposals for putting a GPS-like
> constellation of telecom-relay/clock/positioning satellites around
> Mars.
> 
> In any case, it does not pay to produce the Temporal Coordinates in
> FITS paper until after the current ITU-R decision process calms down
> with some sort of resolution.

That's probably true, although it sounds like people are at least starting to 
think about what it would look like.  I did want to get my $0.02 in early in the 
process, particularly about the distinction between relative and absolute time.

> --
> 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 m
> 
> _______________________________________________
> fitswcs mailing list
> fitswcs at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/fitswcs
> 
> 

-- 
William Thompson
NASA Goddard Space Flight Center
Code 612.1
Greenbelt, MD  20771
USA

301-286-2040
William.T.Thompson.1 at gsfc.nasa.gov