[fitsbits] Re: leap second alert

[William Thompson]

Clive Page <cgp at nospam.le.ac.uk> writes:

>In article <Pine.OSF.3.96.991214080223.23121A-100000 at rlsaxps.bnsc.rl.ac.uk>,
>Patrick Wallace  <ptw at star.rl.ac.uk> wrote:
>>I detect a fair degree of consensus among the views recently expressed
>>in this thread.  I've already sent my comments to USNO, but here's a
>>nutshell version:
>>
>>1) OK, I can see why there's a move to freeze UTC and do away with leap
>>   seconds. The public don't understand them, the time-board manufacturers
>>   never got their act together, and sometimes even observatories haven't
>>   quite tamed them.

>Indeed.  I was astonished to find that ESA still insist on using UTC for
>all their spacecraft operations, when TAI (or even TT) would be so much
>more sensible.  Because they lack confidence in handling leap seconds, it
>turns out that they will be ceasing operations on XMM a little before each
>leap second, and restarting again afterwards.  

This isn't strictly true.  I work on the Solar and Heliospheric Observatory
project which is a joint ESA/NASA mission.  The spacecraft was built by ESA and
operates on TAI time.  The onboard clock generates a six byte time signal
representing the number of TAI seconds since January 1, 1958.  The first four
bytes is the number of seconds, while the last two bytes is the fractional
part, precise to 1/2048th of a second.  (Not all the bits are used in the last
byte.)  All intercommunications between the spacecraft and the scientific
instruments, and between the spacecraft and the ground, are performed in this
TAI time format.

The operations of the spacecraft are performed from the NASA Goddard Space
Flight Center.  On the ground, all operational planning is performed in UTC
time, requiring conversion from UTC to TAI for the actual commanding.  I wrote
some of the software which implements this function within IDL for a number of
the SOHO instrument teams, and keep track of the addition of new leap seconds.
I'm not aware of any concerns regarding operating SOHO through leap second
changes.

>I guess it's true that any fully logical change will cost money, but one
>possible change is just not to bother inserting any more leap seconds after
>the next one.  Slowly UTC will drift away from mean solar time, but
>practicaly nobody will care, for centuries at least.

I believe that this is the actual proposal.  The discussion about "leap hours"
or whatever on the internet is not likely to have any resemblance to what is
actually discussed by the various international organizations.

Do people in general care about whether or not UTC keeps sync with solar time?
I agree that there are few practical reasons why anyone should care.  Daily
sunrise/sunset tables will slowly drift away from their current values over the
centuries.  Tide tables would also drift away from current predictions.  The
formulas for converting local time (as an approximation to UT1) to sidereal
time for amateur astronomical observations would need to take the drift into
account (which will become apparent within decades rather than centuries, I
believe).

However, I believe that there is an emotional component to this argument, above
and beyond the relatively small practical considerations.  Of all the time
formats currently in use, UT1, TAI, TT, etc., the one with the most direct
relationship to everyday life is UTC, which is the standard upon which civilian
time is based.  Religion has been mentioned as one aspect of everyday life
which will be affected by this proposed change.  Many religions in the world
depend upon precise timing of astronomical events, such as sunrises, sunsets,
and lunar eclipses.  (I've been asked a few times for precise times of lunar
eclipses by members of the public who said that they needed the information for
religious purposes.)  If the scientific community abandons the principal of
tracking the Earth's rotation for the time standard currently accepted as the
standard, this may not sit very well in those countries where those religions
hold sway.

There are other emotional considerations beyond religion.  People may not
understand leap seconds, but they do know that there are people out there who
do understand them, and track this information for them.  If we suddenly turn
around and say that something the scientific community has been doing for
almost 40 years is now too difficult and being abandoned, it will put the
community in a bad light with the public.  The public will perceive that the
time that they are using is no longer as precise as it used to be, and will
feel a loss, even if there are little or no practical reasons for that sense of
loss.

A single, precise time standard is absolutely vital for many aspects of today's
civilization, regardless of whether that time keeps track with the Earth's
rotation.  Transportation (e.g. air travel) and financial transactions are two
examples.  It happens that the world has standardized on UTC as the universal
time standard, out of all the possible standards that could have been chosen.
It could have, for instance, chosen TAI as the standard, and we wouldn't have
to worry about leap seconds.  However, society chose to go with UTC, leap
seconds and all.  I worry that if we change the definition of UTC, then it will
put pressure on segments of society to adopt other time standards, and we may
end up no longer having a single, precise time standard accepted all over the
world.

William Thompson