“The Future of UTC Conference” will be held at the University of VA from 29-31 May 2013 in Charlottesville VA, and Harlan Stenn will be presenting a paper there.

For most of our history, we’ve considered a “day” to be a solar day, and a “year” to be the time it takes the earth to rotate around the sun. We’ve even gone to some pains to mark the seasons accordingly, noting the vernal and autumnal equinoxes and the longest and shortest days of the year.

Over the course of history, we’ve noticed that our calendar doesn’t quite match with reality. Starting with the the Julian Calendar (from Julius Caesar in 45 BC), we started accruing error because a Julian year is 365 days’ long. But the actual length is closer to 365.25 days’ long.

The spring equinox originally happened on 25 March. By 300 AD, this had shifted to 21 March, and by 1582, it was happening on 11 March. Pope Gregory XIII created the new Gregorian calendar which incorporated “leap years”, which has pretty much kept the spring and fall equinoxes on their expected dates.

This worked out pretty well, up to the point where we were able to better measure the length of a second, and also the earth’s rotation. We noticed that the rotation of the earth is not constant, and is usually ever-so-gently slowing. There are relatively fixed and known forces that affect the earth’s rotation, primarily the gravitational pull of the moon, and a number of random forces as well, like the paths taken by the jet stream and the Antarctic Circumpolar current. While most of these forces will slow the earth’s rotation, some events, like the 2004 Indian Ocean earthquake, which caused a speedup. The net effect of this is about once every two years we need to add a “leap second” in order to keep the difference between UTC and “mean solar time” less than one second.