Today is longer than yesterday and tomorrow by one second.
The hyperfine transition is the basis of International Atomic Time (TAI). By definition, the outermost electron in an ordinary cesium 133 atom cycles through this transition 9,192,631,770 times in one second.
International Atomic Time (abbreviated TAI after the French Temps Atomique International) began at midnight GMT on the first day of 1958 and has continued advancing forward at the rate of one second every 9,192,631,770 periods of the hyperfine transition in 133Cs. TAI is maintained by the Bureau International des Poids et Mesures (BIPM) in Paris, which periodically averages the time kept by various atomic clocks around the world. The BIPM then disseminates correction factors needed to synchronize these clocks with the master clock in the Observatoire de Paris.
Coordinated Universal Time (abbreviated UTC) is the basis of legal time throughout the world. All local civil times differ from UTC by either a whole number of hours or an odd number of half hours, but never by any other amount. One second of Coordinated Universal Time is the same as one second of International Atomic Time, but UTC and TAI are slightly out of step. TAI marches forward uniformly, while UTC is adjusted from time to time to keep it synchronized with the earth’s rotation.
The earth is not an effective timekeeper. For most of the last two hundred years the mean solar day has been slightly longer than the 86,400 s currently defined by the International System. Universal Time UT, or more specifically UT1, is in effect the mean solar time. It is continuous (i.e. there are no leap seconds) but has a variable rate because of the Earth’s non-uniform rotation period. It is needed for computing the sidereal time, an essential part of pointing a telescope at a celestial source.
When UT1 lags too far behind UTC, a leap second is inserted at the end of the day before January 1 or July 1 as appropriate. When this happens, 23:59:59 is followed by the unusual time of 23:59:60 before turning over to 00:00:00 and starting the next day. In the unlikely event that UT1 were to lead UTC (that is, if the earth’s rate of rotation were to increase) the provision exists for the insertion of a negative leap second. Were this to ever occur, 23:59:58 of one day would be followed by 00:00:00 of the next, skipping 23:59:59 altogether. In any case the absolute difference between UTC and UT1 must never exceed 0.9 s.
|*||Adjustments from 1961 to 1971 followed a more complicated protocol and were omitted. Only the final offset is shown.|
Are leap seconds even necessary?
- In seven or eight centuries the difference between TAI and UT1 will be about an hour.
- By the year 5000, day and night will have reversed; that is, 12 noon will occur in the middle of the night and 12 midnight will synch up with the midday sun.
What would be so wrong with that? Does it really matter what number we assign to a position of the sun in the sky?
In 1582 Pope Gregory XIII managed to extract ten days from the calendar. On 4 October 1582 the Catholic world went to sleep. When they woke up it was 15 October 1582. By 1752 the protestant nation of England and her American colonies also accepted the change. (They needed to add 11 days to catch up.) In 1873 Japan made the switch. (They needed 12 days.) Then Russia in 1917 and China in 1949. (13 days.) The Greek Orthodox Church is possibly the only European agency that has not accepted this change (although the nation of Greece made the switch in 1923).
I think the big thing is that everyone agrees what time (or day) it is. Not that the time is any particular number. Time is a social construct, remember.