legal time

From: Steve Allen <sla_at_UCOLICK.ORG>
Date: Thu, 10 Apr 2003 22:39:45 -0700

Hints about ITU-R SRG 7A indicate that UTC might be altered such that
leap hours replace leap seconds. In a very broad sense this would
satisfy the design that UTC continue to track earth rotation, so in
that sense it could continue to merit a name like the other species of
"universal" time. The leap hour could be incorporated by omitting the
start of daylight saving (summer) time. The necessary legislative
amendment for the aberration could be scheduled well in advance.
Further, because societies around the world are already forced to
endure this sort of legislative meddling with their standard time, it
would not even matter much if the leap hour was implemented by
different jurisdictions in different years.

Presumably the effect of this is to require an omitted daylight change
every time that the value of DUT1 accrues to a half-integral hour. If
we assume that leap seconds are discontinued real soon and we use the
Stephenson and Morrison parabolic fit to delta_T we get the following
schedule for legislative acts of omitting daylight changes:

DUT1 calendar year grace interval
0.5 hours 2603 600 years
1.5 hours 3152 550 years
2.5 hours 3533 380 years
3.5 hours 3844 310 years
4.5 hours 4113 270 years
5.5 hours 4354 240 years
6.5 hours 4574 220 years

The calendar years are, of course, somewhat approximate. I decline to
go farther for fear of extrapolating Stephenson & Morrison's rough
parabola farther into the future than it goes into the past, and
because I don't feel that I share enough with people 100 generations
hence to be sure of what they'll want to do.

On 2000-07-21 (very early in the history of the LEAPSECS list) Rob
Seaman posted a query about the legal issues of disconnecting civil
time and solar time. Subscribers can see that note at
http://rom.usno.navy.mil/cgi-bin/wa?A2=ind00&L=leapsecs&O=A&F=lf&P=667
I now dig further into the scenarios he mentioned.

The proponents of discontinuing leap seconds have given the impression
that legal issues are part of the reason for keeping the name UTC and
not switching to TAI. They seem to want to be able to discontinue
leap seconds in a manner that will avoid involving legislative action.
By a rough calculation it appears that changing UTC from leap seconds
to leap hours pushes off the need to convince any legislatures to take
action for about 600 years. But it really is not as simple as that.

Section 10 of the Metrologia paper by Nelson et al. published in 2001
contains material of the sort presented elsewhere by Levine. It notes
that the legal basis for time in the UK is (naturally) GMT. It also
notes that the legal basis for time in the US is mean solar time of
meridians w.r.t. Greenwich. Effectively, both of these societies
perform their legal actions based on GMT.

So although UTC is, de facto, legal time in many places, it is not
legal time de jure in all places. With leap seconds intact, UTC has
been able to be used as legal time in the UK and the US because the
difference between legal time and UTC is smaller than the
discriminatory ability of most human event timing. (For that reason I
don't even want to consider whether legal time in the UK and US really
means UT2.)

Now contemplate DUT1 at finer resolution. According to McCarthy's
graph, with about 1 leap second per year being omitted from UTC the
value of DUT1 would grow roughly as follows:
DUT1 calendar year
 30 s 2025
 60 s 2050
 90 s 2070
120 s 2085

If UTC switches to leap hours then within 50 years there will be a
full minute difference between UTC and the legal time currently
defined by the UK and the US. When the difference gets larger than
that it will begin to be notable in the time stamps of most events
such as births, deaths, and traffic accidents. Events such as these
which occur around midnight may have large economic significance --
e.g., tax burdens, inheritances, insurance policy expirations.

By the time DUT1 reaches 1 minute there will begin to be court cases
that challenge the validity of timestamps based on UTC. Expert
witnesses will be called in to testify as to how exactly the clocks in
a venue were set, and whether that means the event happened on one day
or another in the legal time of the jurisdiction.

Actually, I expect that the court challenges will happen even sooner.
The increased presence of surveillance cameras means that many more
events are likely to be recorded with accurate timestamps. Therefore
it is conceivable that such court cases could begin to arise when
the value of DUT1 exceeds only a few seconds.

Court cases such as this will force legislatures to consider the
question of the legal timescale. If the ITU-R succeeds in redefining
UTC they will not avoid legislative complications. Instead they will
guarantee legislative complications within their own lifetimes, and
they are aware of this.

There are a lot of ways that this process of international
professional bodies studying the future of UTC might proceed, and
those of us on the outside are at a loss to understand it.
The legislative aspects could be approached either before
or after making a change to the character of civil time, but
upon imagining them the tone of the two seems very different:

    We see overriding reasons to change legal time to atomic time.
    We need to set up a series of legislative changes so that in about
    30 years (maybe sooner?) we can have a big party where the whole
    world celebrates a negative leap minute (or less?) and switches
    civil/legal time from UTC to TAI.

or

    We think you should know that we've stopped adding leap seconds to
    UTC. That means you have about 10 years to legislate a change in
    your definition of legal time. If you don't get it done by then,
    your country becomes lawsuit fodder.

How exactly does this international process work?

--
Steve Allen          UCO/Lick Observatory       Santa Cruz, CA 95064
sla_at_ucolick.org      Voice: +1 831 459 3046     http://www.ucolick.org/~sla
PGP: 1024/E46978C5   F6 78 D1 10 62 94 8F 2E    49 89 0E FE 26 B4 14 93
Received on Thu Apr 10 2003 - 22:39:58 PDT

This archive was generated by hypermail 2.3.0 : Sat Sep 04 2010 - 09:44:54 PDT