Until the middle of the 20th century the Earth was considered as the most precise clock and the second was defined in relation to the mean solar day. In the 1960s came the atomic clocks and a new definition of the second as the duration of a number of transitions in the atom of Cesium. What we call the SI second, SI for Système International.
We knew at that time that the rotation of the Earth was irregular so we choose the second as a mean value. It just happened that from the introduction of the SI second the Earth rotation started to slow down probably more that it was anticipated implying that time measured with atomic clocks (what we call UTC) started to drift away from mean solar time.
In the 1970’s the mean solar day was about 1 ms longer that the UTC day. It’s a very short time but after a year the difference is about 1 s.
In 1972 it was decided to implement a procedure to adjust UTC with mean solar time. Any time the difference between UTC and UT1 or mean solar time was going to approach 0.9 s there would be either the insertion or the deletion of a second. What we call the leap second.
The last adjustment was in June 2012, as of today the difference between UTC and UT1 is 0.6 s and it was decided to make a further adjustment this year so the last day of June will be 1 s longer than usual and a leap second will be inserted at midnight UTC.
[NOTE: Over on Twitter, our friend Michael Cope, who has been working on leap second compatillity, says this]:
@littleatoms @mePadraigReidy fact check: Last leap second was December 31 2012 not June. Also happened to be a Saturday so markets closed.
— Michael Cope (@miketcope) June 30, 2015
As our conception of time is concerned this will not affect us, adding a leap second in a way is similar to the extra hour for daylight saving except that if you can sense the difference between the 24 hours day and the next 23 or 25 hours. But you won’t feel any difference with a day lasting 86,400 seconds and the next one lasting 86,401 seconds.
However the matter is different for digital systems. These systems are highly dependent on very precise time synchronization based on UTC so any time there is a leap second you have to make sure that it is inserted correctly and accepted worldwide. As leap seconds are only announced a few months in advance this cannot be done automatically, you need manual intervention and verification to make sure things are done properly and there won’t be any mishaps in the process. There were some problems in 2012 with internet sites and aircraft booking to name a few.
UTC time as we know it today is quite a recent invention dating from the middle of the 20th century, and as the second is constant it is not linked any more to the Earth rotation. In fact before 1967 we adjusted our clocks with solar time and now we need to do the opposite adjust the Earth rotation with UTC. Unfortunately as there is no way to make the Earth rotate quicker or slower we decided maintain an artificial link to the Earth rotation through the implementation of the leap second.
If we stop inserting leap seconds we break that link and UTC and UT1 will start to diverge from each other. There were 25 leap seconds inserted in 40 years so we can expect that divergence to be at least 1 mn per century. That is causing some concerns for some administrations.
Another concern is for systems that require UTC close to UT1 such as astronomical observatories or celestial navigation, they will need to be updated to adapt to the new time and been able to recover UT1.
Finally another group of administrations is of the opinion that stopping the insertion of the leap second will be beneficial for the digital world where we live. You have to realize that there were problems with almost every leap second in the last 40 years although not as important and publicized as in 2012. Stopping the insertion of the leap second will mean saving time and money on the software update and removing the risks associated with inserting the leap second.
