Farewell to the Leap Second

In a few days we will reach the December solstice, with the shortest hours of sunlight in the northern hemisphere (a few seconds less than 9 hours and 35 minutes in Lynchburg, Virginia) and the longest days in the southern hemisphere (more than 14 hours and 12 minutes in Port Macquarie, New South Wales, Australia). The exact time of when the southern hemisphere of Earth is most tilted toward the Sun is 4:48 p.m. EST on December 21st, or 21:48 UTC on the same date.

So who keeps track of this stuff? How do we tell time, anyway?

There are multiple time standards, but here we will limit ourselves to just three, and how they define a single day.

  • Astronomical time: based on the repetition of the day/night cycle; one day is the time from one noon to the next, or from one midnight to the next
  • UTC (Coordinated Universal Time): Primary time standard for the world; one day is within one second of the mean (average) length of an astronomical day
  • International Atomic Time (TAI): a weighted average of the time kept by over 450 atomic clocks in over 80 national laboratories worldwide. These employ cesium atoms, and the second is defined as “the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium-133 atom”

Did we lose you with that last one? Stay with us.

Astronomical time was more than sufficient for most of humanity’s existence. As more precise timekeeping devices came into use, our ability to divide and measure the day in ever smaller units followed in their wake. There were 24 hours in the day, 60 minutes in an hour, and 60 seconds in a minute, numbers derived from the sexagesimal notation of Mesopotamian civilizations.

But the astronomical day is not always exactly the same length. It varies throughout the year because of the varying speed of the Earth around the Sun, by as much as 20 seconds It can shift with earthquakes and other geological events Even weather events can change it. We mostly get around this by defining the mean solar day, and over the course of a full year, that doesn’t change much.

Except…the Earth’s rotation is slowing down. The gravitational pull of the Moon means that the length of the mean solar day is getting ever so slightly greater.

https://qph.cf2.quoracdn.net/main-qimg-9edeca356f46dbdd84e466f133428d18-lq

So let’s jump ahead to those atomic clocks. They are incredibly precise. They can measure time intervals as small as a nanosecond—a billionth of a second! They can measure the difference in the rate of time passage between different locations on the Earth’s surface. Gravitational time dilation causes clocks at lower altitude to run more slowly. TAI is therefore an average weighted to give the time at sea level, absent the effects of wind or tides. It is continuous, no adjustments for astronomical variations.

The time that the world sets its clocks by is UTC, which is effectively what we used to know as Greenwich Mean Time. It relies on the astronomical definition of a day, which is set at 86,400 seconds (60 seconds/minute x 60 minutes/hour x 24 hours/day). But more precise atomic measurements show that the length of the mean solar day is actually longer than this. Over time, that disparity accumulates.

The solution adopted in 1972 was to add a leap second to UTC whenever it fell out of sync with TAI by that much. This happens roughly every 800 days. There have been 27 leap seconds inserted since 1972; the last one was on December 31, 2016.

The modern world of high speed computing in automation and in trading systems makes leap seconds problematic. The international body responsible for such things has proposed eliminating the leap second by 2035. At some point around 2100, UTC will be a minute out of sync with astronomical time. Perhaps the final minute of the day will stretch to two minutes. Mark your calendar so you can use the extra time wisely!

 

 

Tagged with: , ,

Leave a Reply

Your email address will not be published. Required fields are marked *

*