Astrophysics (Index)About

leap second

(one-second adjustment to make standard times track solar time)

A leap second is a second added to or subtracted to some standard time to keep that time standard aligned with the Earth's actual rotation period. In particular, such adjustments have been made to UTC to keep it from drifting away from UT1 the former which counts SI seconds, with 86400 for a day, and the latter which counts time by measuring Earth solar days. The problem is that Earth's rotation varies, and these two get out of sync. The variation is small and erratic, but can be large enough that a discrepancy of a second can occur over the course of a year. Variation is due to weather, ocean currents, ice movement, tidal forces, and Earth activity (e.g., plate tectonics, volcanos, and movement of material below Earth's crust).

For astronomical purposes, a time standard that merely counts standard-length seconds is most useful: to derive the precise length of time between two observations that are years apart (i.e., precise to a second or less), it is pretty-much required to use a standard-length second (such as the SI second), and a time system that simply counts them, 86400 per day. Some decades have had more than one leap second, and some have had none. Since the adjustments are carried out based on ongoing measurement, it is not clear years in advance when an adjustment will be made.

But for normal, everyday time, keeping noon as the exact time when the Sun is at its zenith over a particular meridian seemed reasonable, and formal leap seconds were introduced in 1972, to be added when needed at the end of June 30 and/or December 31. By 1971, UTC was about 10 seconds different than UT1, so starting in 1972, seconds were added once or twice a year to close this difference, and then seconds were added to keep the difference to less than a second. The first was added at the end of June 30, 1972.

A revision is planned for UTC, that leap-second adjustments will no longer be made: in the future, UTC will be allowed to diverge from UT1 by up to an hour (3600 seconds), at which time a leap hour will be added or subtracted. How long it will take to reach the first such leap hour depends upon the erratic changes in Earth rotation, but it will certainly not occur for hundreds or even thousands of years. The official date to change UTC in this manner is May 20, 2027. There hasn't been a leap second in about a decade (the most recent was at the end of December 31, 2016, which was the 27th leap second), so I suppose we can expect zero or one between November 2026 and May 2027.

The reason for the change is that computing, networking, and a lot of technology depends upon second and sub-second timing, and a leap second adjustment ends up basically requiring a software update for each kind of machine deployed to each instance. Changing all the "clocks" is a logistical nightmare, and even coding the programming logic to adjust the clocks is a complex requirement, often leading to bugs. Astronomy doesn't use UTC for its precise timing: Terrestrial Time is its preferred time system for many purposes, which hasn't had leap seconds.


(time)
Further reading:
https://en.wikipedia.org/wiki/Leap_second
https://dictionary.obspm.fr/terms/leap-second/
https://www.nist.gov/pml/time-and-frequency-division/leap-seconds-faqs
https://www.scientificamerican.com/article/scientists-vote-to-end-leap-seconds-redefining-how-the-world-keeps-time/
https://www.cnmoc.usff.navy.mil/Our-Commands/United-States-Naval-Observatory/Precise-Time-Department/Global-Positioning-System/USNO-GPS-Time-Transfer/Leap-Seconds/

Referenced by page:
Julian date (JD)

Index