[LEAPSECS] the epoch of TAI, with no more doubt

[Steve Allen]

On Sun 2019-01-20T06:22:58-0800 Tom Van Baak hath writ:
> I'm curious how your findings compare with this random link I ran across [1]:
>
> https://koka-lang.github.io/koka/doc/std_time_utc.html
>
> See especially section "1.3. UTC before 1961"

It was 1959 July or August (recollections differ)  when H.M. Smith of
the Greenwich time department had some UK and US folks in his living
room for tea.  The UK and US were already independently regulating
their time broadcasts using cesium, with USNO and US NBS using
different strategies.  That tea-time was the impetus for coordination,
and it was following a recommendation for frequency stability from the
1959 CCIR meeting in Los Angeles which basically said that VLF
transmitters had to be calibrated using cesium.

USNO Time Service Notice no 9 (1960-12-21) says that US and UK started
coordinating in 1959, and the author of that was at the tea-time
meeting.  Without transcribing and plotting the BIH records it is hard
to say when exactly the US and UK coordination started, but one
memoire says the official US/UK agreement was forged during 1960.  It
is clear that during 1960 the URSI meeting recommended that every
transmitter should coordinate using cesium, and that BIH should have
the job of choosing the coordination starting with 1961.

Really there can be nothing called UTC before 1960.  Everything before
then in that github project applies to one particular source of
broadcast time signals, or is a wishful reconstruction based on no
actual source at all.  And as I noted recently, everything in that
github between 1970-01-01 and 1972-01-01 does not apply to anyone in
Europe who was getting time stamps using DCF77 after they decided it
was illegal to broadcast rubber-second old UTC.

> and also the link:
> https://books.google.com/books?id=uJ4JhGJANb4C&lpg=PA87&vq=wwv&pg=PA87#v=onepage&q=wwv&f=false

USNO Time Service Notice no 8 (1959-11-18) disagrees and says that for
1959 the USNO cesium offset was -170e-10 and for 1960 the offset was
-150e-10.  This is not the only place where contemporary documents
differ from McCarthy and Seidelmann and their books.

All of the USNO broadcast peculiarities and changes were indicated in
the circulars issues by the time division.  A complete collection of
the USNO time division circulars may not exist because most recipients
did not save them, and in any case the acid of the paper they used has
turned most of them to dust.  Usually the circulars served as
announcements in advance so that navigators would not be surprised by
changes, and usually the circulars give the rationale for the changes.

Time signal broadcasts have always been extrapolations of time based
on the local clocks at the station and algorithms for that
extrapolation which have particular strategies.  According to Bulletin
Horaire WWV was using a different strategy than USNO until the era
when coordination began.

Before coordination most observatories kept an internal astronomical
time scale and noted the difference between that and the radio
broadcast time signals that they controlled.  Bulletin Horaire shows
how the UK broadcasts were providing Provisional Uniform Time
regulated using quartz by the early 1950s, and very clearly shows when
the Greenwich observatory decided to change their official internal
time from the direct astronomical observations to the readings from
the quartz chronometers.

In the very old Bulletin Horaire from around 1930 it is noted that WWV
was particularly good for the purposes of geodesy.  Up until the era
of coordination the USNO broadcasts had a systematic difference from
other broadcasts because USNO was for navigators.  USNO broadcast time
that was self-consistent with longitudes on the navigational charts,
and those charts were based on an old value for the longitude of
Washington.  USNO Time Service Notice no 9 (1960-12-21) indicates that
on 1961-01-01 they changed the time at the USNO observatories by 30 to
40 ms in order to fix the longstanding time offset between the USNO
observatories and the global average notion of longitude.

> Give it some thought.  I don't have a quick answer and would
> appreciate your instr.  If you've ever worked with raw UTC(k) data,
> present or historical, you know there's no one right time.  So in the
> late 1950's, making claims about Paris vs.  Washington / Gaithersburg
> / Boulder might not be as clear as you hope.

> In my home museum the earliest cesium standard is 9,192,631,840 Hz
> not 9,192,631,770 [2].  I've learned that re-interpreting
> international time & frequency history is not always simple.

Almost every writeup of the history of time ignores details which are
evident in the issues of Bulletin Horaire and the bulletins issued by
the various national time service bureaus.  The dirty laundry of
timing for all the received radio signals is there in Bulletin
Horaire.  There are hundreds of thousands of turgid numbers giving the
timing of exactly when each signal was received, and then followup
publications indicating how wrong each station clock was.

Some of the discussions in Bulletin Horaire make references to changes
made by observatories and transmitters, but most of the particular
decisions and rationale made by the folks doing the broadcasting are
lost to history.  Transcribing and plotting Bulletin Horaire would
allow for the reconstruction of a lot of when changes happened and
good guesses at what the underlying strategy was.  And, alas, reading
through Bulletin Horaire also calls into question some content in the
historical writeups.

I will admit that it borders on insanity to actually read through
Bulletin Horaire and pick out details.  I suppose Guinot must have
agreed, and I think that is why the precise records of earth rotation
that the IERS inherited from the BIH begin at the transition from FK3
to FK4 on 1962-01-01.

--
Steve Allen                    <sla at ucolick.org>              WGS-84 (GPS)
UCO/Lick Observatory--ISB 260  Natural Sciences II, Room 165  Lat  +36.99855
1156 High Street               Voice: +1 831 459 3046         Lng -122.06015
Santa Cruz, CA 95064           https://www.ucolick.org/~sla/  Hgt +250 m