Does Anybody Really Know What Time It Is?

Title: The Future of Time: UTC and the Leap Second

Authors: David Finkleman, Steve Allen, John H. Seago, Rob Seaman, P. Kenneth Seidelmann

First Author’s Institution: Center for Space Standards and Innovation

What does a day in the life of a human mean? Many philosophers have spent lifetimes pondering this question and the ontological threads that branch from it. There is not enough time to go into the conversation here, but we can still reflect on how the quantitative nature of the day has changed over the course of human history.

Since antiquity, the concept of a day has perhaps always been the most fundamental, measured by the time it takes the sun to transit the meridian, set, rise, and return to the meridian. The month was a longer period of time that relied upon the phases of the Moon (perhaps we should call it a moonth…). The year measures the time it takes the Earth to orbit around the sun, or–if you weren’t yet a card-carrying member of the heliocentric school of thought–the time it took for the sun to return to the same place against the constellations. A favorite device of sci-fi writers is to create extra-terrestrial societies with disparate years, months, and days that differ from those of Earth (and which would be entirely realistic). None of these units of time depend explicitly on the other, yet we use them to segment our lives into units of time, and because they do not fit perfectly into each other, the remainder of the division is the domain of leap years.

The units we use to measure time have until recently been always derived from the movements of the heavens, which we have considered immutable because we have never had the accuracy of measurement to notice otherwise. Starting with the Babylonians, we divided our days into neatly divisible blocks of hours, minutes, and then seconds. The advent of extremely accurate atomic clocks that rely upon the hyperfine quantum transitions of Cesium-133 atoms have challenged our traditional notions of time.

Now might be a convenient time to begin watching the accompanying World Wide Telescope tour below. If you are reading on an RSS reader, go here:

If we use a sundial to measure the time of day, we would be measuring the apparent solar time (which is not uniform throughout the year). Once you’ve watched the tour, consider checking out this neat interactive demonstration of analemmas by Wolfram.

However, if we were going to stretch this idea of the apparent solar day to be consistent over the year, we would reach the current idea of the mean solar day, which is our normal concept of the day as measured by our wristwatches.

Astronomers have long used the concept of a sidereal day, which is the time it takes for the Earth to make one rotation against the background stars. As seen in the tour, because the Earth moves in its orbit around the sun, a sidereal day is only 23 hours and 56 minutes long.

The onset of atomic timekeeping has created a situation that is untenable and unsustainable. Clocks are so accurate that we can now detect the slowing of the Earth’s rotation. Using these clocks, we can define a uniform time tied to atomic (and presumably constant) processes rather than to dynamic processes like the Earth’s orbit around the sun or the Earth’s rotation, which we know can easily change over long periods of… time.

An analemma over the Ukraine (credit APOD). This image was formed by stacking images taken at the same time every day. Notice how the sun's position in the sky changes over the course of the year.

In their article in the July-August 2011 issue of American Scientist¬†(posted to the arXiv), Finkleman et al discuss the nature, measurement, and future of our concept of time. Because the accuracy of atomic clocks make the drift in the Earth’s rotation period apparent (due to the Moon and lunar tides) corrections are needed if we wish our clocks to be synchronized with the motions of the sun. It wouldn’t make much sense if 11pm actually corresponded to what we now consider noon. Thus, since 1972, leap seconds have been occasionally added to keep these measurements synchronized.

As one can imagine, adding these leap seconds to every clock in the world is not straightforward. One would hope that a normal person wouldn’t mind if their wristwatch were only a second off (in fact, most of us might wish for that much accuracy…) but our GPS satellites and communication networks care very much. Sub-second accuracy is essential to their functionality. So how do we all synchronize these clocks? Not every timekeeping device is capable of just inserting a leap second (11:59:60 pm) on command.

An organization affiliated with the United Nations, the Radiocommunication Sector of the International Telecommunications Union (ITU-R), has proposed that we eliminate the difficulty of leap seconds all together–let the clocks diverge from the heavens!

Tweaking calendars is not a new practice conjured up for the atomic age. Institutions like churches, governments, and emperors have changed their calendars many times to account for errors in time keeping. Regularly we join the likes of the Pope Gregory XIII, Sweden, and Julius Caesar!

The World Wide Telescope (WWT) tour was created by Katherine Rosenfeld and Ian Czekala. Microsoft’s telescope software lets you interactively explore the universe. We’ve also entered our WWT tour in the Summer WWT Tour contest. If you’d like to give it a shot, you can make your own tour too.

About Ian Czekala

I am a second year graduate student at the Harvard-Smithsonian Center for Astrophysics. I work with Edo Berger on studying intermediate luminosity optical transients discovered with Pan-STARRS.


  1. We have a telescope by the same manufacturer as Pan-Starrs, and we know it will fail to point correctly if leap seconds are abandoned in UTC.

  2. Nice work astrobiters! Now I want an animated analemma in WWT… there’s a challenge!

  3. A correction: It was the July-August 2011 issue of American Scientist, not Scientific American.

    A clarification: The question of continuing leap seconds is not the same thing as the actual question up for a vote at the ITU of redefining Coordinated Universal Time. The consensus of the only public meeting (Torino, 2003) held to date was that should leap seconds be ceased, the resulting timescale should be given a different name. This would preserve UTC as a source of Universal Time. The debate isn’t about improving access to atomic time (already available as TAI and GPS), it is about eliminating access to Universal Time. The two kinds of time are distinct and will remain different things whatever the ITU claims.

    An observation: The sidereal day is the actual day on Earth – the period needed to rotate 360 degrees. The best way to understand what the phrase “mean solar day” means is to realize it is simply the sidereal day adjusted by one (the integer “1”) day per year due to lapping the Sun annually.

    An opportunity: Another public meeting ( will be held 5-6 October 2011 near Philadelphia. The issues are not clear cut. The implications dire for astronomical telescopes, systems, and software. For example, WWT explicitly assumes UTC will remain Universal Time whenever you run the solar system clock forward in time.

    Precis: This is an artificial crisis. Consensus should be sought before taking action.

  4. Steve, Alyssa, and Rob, thank you all for your comments and correction. We greatly appreciate your feedback!


Leave a Reply