About Time: Is the June 30th Leap Second the Last?

Article written: 23 Jun , 2015
Updated: 16 Sep , 2016

The month of June 2015 is just a tad longer than usual… but not for the reason you’ve been told.

Chances are, you’ll soon be hearing that we’re tacking on an extra second to the very end of June 30th, though the reason why is a bit more complex than the explanation you’ll be hearing.

It’s an error that comes around and is repeated about every 500 days or so, as we add a leap second to June 30th or December 31st.

‘The rotation of the Earth is slowing down,’ your local weather newscaster/website/anonymous person on Twitter will say. ‘This is why we need to add in an extra second every few years, to keep our accounting for time in sync.’

Image credit:

The observed variation of the Earth’s rotation in milliseconds since the adoption of the leap second. Image credit: The United States Naval Observatory

Now, I know what you’re thinking.

Doesn’t adding a second once every 18-24 months or so add up to an awful lot? Are we really slowing down to the tune of (calculator apps out) over 11 minutes per millennium? What’s going on here?

Here’s what your weatherman won’t tell you.

The story of the second and the insertion of the modern day leap second is a curious case of modern astronomical history.

Universe Today recently covered the quirks of the Earth’s rotation on this past weekend’s June solstice. We are indeed slowing down, to the tune of an average of 2.3 milliseconds (thousands of a second) of a day per century in the current epoch, mostly due to the tidal braking action of the Moon. The advent of anthropogenic global warming will also incur variations in the Earth’s rotation rate as well.

Historically, the second was defined as 1/86,400th (60 seconds x 60 minutes x 24 hours) of a mean solar day. We’ve actually been on an astronomical standard of time of one sort or another for thousands of years, though it’s only been over the last two centuries that we’ve really needed—or could even reliably measure—time to an accuracy of less than a second. These early observations were made by astronomers using transit instruments as they watched stars ‘cross the wire’ in an eyepiece using nothing more sophisticated than a Mark-1 eyeball.

Image credit:

A transit instrument on display at the Quito Observatory in Quito, Ecuador. Image credit: David Dickinson

The whole affair was addressed in 1956 by the International Committee for Weights and Measures, which defined what was known as the ephemeris, or astronomical second as a fraction—1/31,556,925.9747th to be precise—of the tropical year set at noon on January 1st 1900.

Image credit:

Simon Newcomb. Image in the Public Domain

Now, this decision relied on measurements contained in Simon Newcomb’s 1895 book Tables of the Sun to describe the motion of the Earth. Extrapolating back, a day was exactly 86,400 modern seconds long… in 1820.

In the intervening 195 years, the modern day is now about an extra 1/500th (86,400.002) of an SI second long. In turn, the SI second was defined in 1967 as:

The duration of 9,192,631,770 periods of radiation corresponding to the transition between two hyperfine levels of the ground state of the Cesium-133 atom.

Image credit:

An atomic clock at the Federal Office of Metrology in Bern, Switzerland. Image credit: Wikimedia Commons/Public Domain

Now, physicists love to have an SI definition that isn’t reliant on an artifact. In fact, the pesky holdout known as the kilogram is the last of the seven SI base units that is based on an object and not a constant that anyone can measure in a lab worldwide. Simply locking a second at 1/86,400th of a mean solar day would mean that the second itself was slowly lengthening, creating its own can of worms…

So the leap second came to be, as a compromise between UT1 (Astronomical observed time) and UTC (Coordinated Universal Time), which defines a day as being comprised of 86,400 SI seconds. These days, the United States Naval Observatory utilizes observations which include quasars, GPS satellites and laser ranging experiments left on the Moon by Apollo astronauts to measure UT1.

The difference between Universal and Terrestrial Time is often referred to as Delta T.

Image credit:

An 1853 Universal Dial Plate depicting time worldwide before the adoption of Universal Time. Image credit: Wikimedia Commons/Public Domain image

The first leap second was inserted on June 30th 1972, and 25 leap seconds have been introduced up until the extra June 30th second next week.

But the Earth’s rotation isn’t actually slowing down a second every time we add one… this is the point most folks get wrong. Think of it this way: the modern Gregorian calendar inserts a leap day every four years to keep it in sync with the mean tropical year… but the length of the year itself doesn’t increase by a day every four years. Those fractions of a second per day just keep adding up until the difference between UT1 and UTC mounts towards one second, and the good folks at the International Earth Rotation Service  decide something must be done.

And don’t fear the leap second, though we’ve already seen many ‘Y2K redux’ cries already cropping up around the web. We do this every 18-24 months or so, and Skynet hasn’t become self-aware… or at least, not yet.

Of course, programmers hate the leap second, and much like the patchwork of daylight saving time and time zone rules, it causes a colossal headache to assure all of those exceptions and rules are accounted for. Consider, for example, how many transactions (emails, tweets, etc) fly around the globe every second. Many services such as Google instead apply what’s known as a ‘leap smear,’ which slices the leap second out into tinier micro-second sized bites.

With the current system in place, leap seconds will become ever more frequent as the Earth’s rotation continues to slow. There have been calls over the years to even do away with the astronomical standard for measuring time entirely, and go exclusively to the SI second and UTC. This would also create a curious situation of not only, say, throwing off local sunset and sunrise times, but users of GOTO telescope pointing systems would probably note errors within a few decades or so.

This coming November, The World Radiocommunication Conference being held in Geneva, Switzerland is looking to address the issue, though we suspect that, for now at least, the future of the leap second is secure… perhaps, if we did indeed go off the astronomical time standard for the first time in the history of modern human civilization, a leap hour might have to be instituted somewhere around oh say, 2600 AD.

What do you, the reader think? Should it be ‘down with the leap second,’ or should we keep our clocks in lock step with the cosmos?

, , , , ,

2 Responses

  1. Change the count of the Cesium-133 atom so a leap second only needs to be added every 20 years or so instead of every 500 days.

  2. chfosmith says

    Hi David
    Your comment:
    “The advent of anthropogenic global warming is also increasing this slowdown by a small factor as well.”
    is not correct, and is the same mistake made by Becky Nevin, whom you referenced.
    As the glaciers melt on Greenland, in the Swiss Alps, or Kilimanjaro, the conservation of momentum will speed up the rotation of the earth.
    When the pack ice melts in the Arctic, the centre of mass of the previous ice moves towards the centre of the earth with the same slight speed up effect,
    unless the ice is exactly at the North Pole.
    Note: When the Arctic ice melts, no seawater is displaced to the equator, the mass remains the same.
    Simple experiment: Put icewater in a glass. When the ice melts, the water level in the glass does not change.
    To be more exact, put a lid on the glass so there is no evaporation of the water or sublimation of the ice.


Comments are closed.