Leap Seconds

Subscribe
Apple | Spotify | Amazon | Player.FM | TuneIn
Castbox | Podurama | Podcast Republic | RSS | Patreon


Podcast Transcript

Every few years, without any regular schedule or planning, officials at the International Earth Rotation and Reference Systems Service will add an extra second to the year. 

The reason why this is done does make sense, but adding an extra second to a year can cause a host of problems, and many are wondering if it is a practice that should be continued. 

Learn more about Leap Seconds, why they exist, when they happen, and if they should continue to exist on this episode of Everything Everywhere Daily.


This entire episode is about one second. One second which may or may not be tacked on to a year. 

One second over the course of a year might not seem like much, and in a human frame of reference, it isn’t a big deal. If it weren’t for the fact that occasional leap seconds were publicly announced, no one would ever know they even occurred. 

Before I get into why leap seconds exist, it is important to know just what a second is and how it is defined, because it is the key to understanding everything I’ll be discussing.

I’ve done several episodes on the subject of time. I’ve talked about why the calendar is the way it is, how the months were named and the order they are in, and why there are seven days in a week. 

However, the fundamental unit of time is not a year, month, week, or day. It is the second. 

Defining a unit of time is much more tricky than defining a physical unit. When the meter and kilogram were defined, they were defined based on an actual physical object which sat in a vault in Paris

There was a literal piece of metal, and that piece of metal was, by definition, one meter. The same was true with a metallic weight which was the definition of a kilogram.

You can’t quite do that with a second. 

A second is a unit of time so short that the ancient really didn’t have any use for it. There was no way to track time with such precision, and there were no uses for a unit of time so small.

The modern second was created around the year 1000 by the Islamic scholar Abu Rayhan Muhammad ibn Ahmad al-Biruni. 

Al-Biruni codified the system of dividing a day into 24 hours and then dividing an hour into 60 minutes and a minute into 60 seconds. 

So, by al-Biruni’s definition, a second was 1/86,400th of a day. 

For all practical purposes, this is still the definition of a second that we all use today. 

When the metric system was codified in the 19th century, the base unit of time was the second, and it used this very definition. 

However, there were problems. 

As scientists developed more precise instruments, they discovered that the length of a day could vary every so slightly. 

Because the length of a day wasn’t absolutely consistent, there needed to be another way to define a second. 

The next logical step was to simply define a second in terms of a year. 

The length of a year shouldn’t have anything to do with what is happing to the Earth’s rotation. 

So, in the 1950s, the new definition was that a second was 1?31,556,925.9747th of a year. 

But that, too, had issues and wasn’t accurate enough to be of use for the increasingly precise needs of researchers and new electronic technologies.

So instead of using an astronomical definition of a second, they went in the other direction and used an atomic definition of a second. 

The definition they selected was that a second was the time it took for 9,192,631,770 vibrations of a cesium-133 atom. Cesium was chosen simply because it has one electron in its outer electron shell, which made it very easy to observe. 

This definition of a second using cesium atoms was very close to the astronomical definition, to within one in ten billion. 

This is still the definition of a second that is used today, and it is a definition that works well and has been able to work with even more incredibly precise atomic clocks.

While decoupling the definition of a second from astronomical events such as a day or a year made it much more accurate, there was still the problem of days and years. 

As I’ve mentioned many times in previous episodes, one of the reasons why some time measurements are so confusing is that time units often don’t divide into other time units evenly. 

In the case of seconds and years, here, too, with the new definition of a second, it wasn’t perfectly divisible by a year anymore. It was really, really close, but it wasn’t perfect. 

So just like an extra day is necessary once every four years to keep calendars in sync, it is necessary every so often to keep things in sync. 

Generally speaking, there are two different systems that are used to keep track of time. One is called “Universal Time,” which is based on the rotation of the Earth. The other is called “Coordinated Universal Time” or UTC which is based on the definition of a second from cesium atoms. 

The leap second is used to ensure Universal Time and Coordinated Universal Time don’t get too far away from each other. 

However, there is more to it than that. If it was simply compensating for the differences in the times, leap seconds are something that could be scheduled out in advance, just like a leap year. 

However, that isn’t the case. 

That is because the rotation of the Earth can change. 

One change is a steady change that slows down the rotation of the Earth over long periods of time. This is due to what is known as tidal friction from the moon.  The effect is about 2.3 milliseconds per century.

Tidal friction acts very slowly. So slow that it will take 200 million years for the Earth to have a day that is 25 hours long. Likewise, when dinosaurs walked the Earth, a day would have been 23.5 hours long.

Tidal friction acts pretty steady and slow. 2.3 milliseconds a century wouldn’t be enough to warrant a leap second in our lifetimes. 

There is something else. Events on and in the Earth can change the Earth’s speed of rotation.

This mostly comes from earthquakes and shifts in crust and mantle.  Massive shifts in rock can change our planet’s moment of inertia, which can slightly change the rate of spin.

This is similar to how figure skaters can spin faster by pulling in their arms or slower by extending their arms. 

For example, when land that was formerly covered by ice age glaciers began to rebound, it slightly increased the speed of the Earth’s rotation by 0.6 milliseconds per century. 

Likewise, the great 2004 earthquake that caused massive tsunamis around the Indian Ocean increased the rotation by 2.68 milliseconds per century. 

Other similar events might tend to slow rotation by small amounts. 

The end result is that every so often, the clocks need to be adjusted to make sure they are in sync. 

The problem is because it is so influenced by events on Earth, you can’t predict when a leap second will be necessary. 

The first 10 leap seconds were counted in 1972. Since then, there have been 27 times when a leap second was added. 

The seconds are added to the UTC clock, which is man-made, not the UT clock, which is determined by the actual rotation of the Earth.

The way a leap second is added is by adding it to the last second of the year on December 31 or by adding it to the last second of the first half of the year on June 30. 

For example, normally, in the count up to midnight, at 11:59 pm, the seconds would count 57, 58, 59, 00, with 0 being midnight and the start of the new day. 

For a leap second, the count would go, 57, 58, 59, 60, 0.  

The second is added at the same time everywhere around the world as UTC is the same everywhere in the world.  Where I live, the time in December is -6 UTC, so the leap second would occur 1 second before 6 pm on December 31.

Leap seconds historically have happened about once every two years. However, the last leap second was in 2016, and from 1972 to 1979, there was at least one leap second every year. 

The IERA or International Earth Rotation and Reference Systems Service determine if a leap second is needed.

Their job is to ensure that the difference between UT and UTC is never more than .9 seconds.  If the clocks should differ by more than that, they will announce a leap second, usually about six months in advance.

So far, so good. We introduce leap seconds to keep the Earth’s clock in check with our atomic clocks. 

However, there is a problem with leap seconds. Getting all of the world’s clocks to recognize it is a massive pain. This has become a greater issue as we rely more and more on precise timekeeping. 

Financial markets, computer networks, and cellular networks, all require subsecond coordination of time. 

The infrequent and irregular nature of leap seconds is what creates the problem. Leap seconds can also occur in the middle of the business day in Western North America and East Asia. 

One of the solutions being floated is to just get rid of leap seconds. If you think about it, it wouldn’t really make a difference if the UT and UTC started to diverge by seconds. 

Time zones have already made it such that noon doesn’t actually occur at the sun’s zenith in most places.  If we just ignored leap seconds, it would take thousands of years for the clocks to diverge by even 1 hour. 

It might be easier every 200 years to do a leap minute, which could be planned well in advance, than it would be to continue doing irregular leap seconds. 


There are time systems right now that are in heavy use, which don’t have to worry about leap seconds.

UNIX time is a time system used by computers that doesn’t bother with years, months, days, hours, or minutes. It just counts the number of seconds that have passed since midnight, January 1, 1970. That’s it. It is just one massively long number. There is no need for leap seconds because the clock doesn’t sync with anything. 

The current UNIX time as I am writing these words is 1,668,681,423. 

Another well-used time system is GPS time. The American global positioning system doesn’t use leap seconds, and it is incredibly accurate. 

The United Nations’ International Telecommunication Union has proposed just adopting GPS time instead of UTC. The main objection is that the GPS system is run by the US Space Force, and most countries prefer UTC, which has some sort of international oversight.

GPS time is currently 18 seconds ahead of UTC. The Chinese navigation satellites are 4 seconds off of UTC.

Other navigation satellite systems do use leap seconds, in particular, the Russian GLONASS system. 

To really throw a wrench into things, it is looking like it might very well be necessary, in about seven years, to have a negative leap second. 

The organization which oversees UTC is the International Bureau of Weights and Measures. They are the group that will determine if leap seconds will be continued in the future for UTC.

Right now, getting the world on the same page in terms of timekeeping is probably more important than keeping UTC synched to the rotation of the Earth by a few seconds. 

We’ll still know what the actual time of the rotation of the Earth is, we just wouldn’t need to track it precisely for computerized transactions and communications. 

Depending on what the International Bureau of Weights and Measures decides and when they decide it, it might be very possible that we have witnessed our last leap second.