Time On Planet Earth

But then, of course, we want to divide time up into smaller units (hours, minutes, seconds). But solar time gives us a problem with that because the interval from the instant the Sun is overhead on one day until the instant the Sun is overhead on the next day is not a constant; it varies over the course of a year, by about 15 minutes. (This is because the orbit of the Earth around the Sun isn't a perfect circle (it's an ellipse) and because the Earth's rotational axis is not at an exact right angle to the plane its orbit is in.)

So that's not quite an adequate definition of a "day".

The next refinement is mean solar time -- "mean" meaning "average". Mean solar time is defined as almost the same as solar time (i.e. calling it noon when the Sun is directly overhead), except we average it out so that the interval between "noon" on one day and "noon" on the next is always the same. Then we can divide this into 24 hours, etc.

Note that mean solar time changes as you move east or west. When the Sun is directly overhead here, it's midnight on the other side of the planet. When the Sun is directly over the Bahen building, that's about two minutes before the time the Sun will be directly over downtown Hamilton (Ontario).

Now, in the days when a trip from Toronto to Hamilton would take you all day, and your pocket watch was very inaccurate by modern standards anyway, that two minutes isn't a big deal. Nevertheless, the fact that the mean solar time here is a fraction of a second different from the mean solar time in the Galbraith building across the street seems downright silly. Pretty early on in the history of clock-making, people adopted the idea of standard time for an entire town or city -- the standard time might be set by a church with a big clock tower, for example. We'd consider the time to be the same in the entire city.

Nowadays, when we can take a train from Toronto to Hamilton in 30 minutes, and our wristwatches are accurate to the second, being off by two minutes when you get there seems almost as silly as the idea of having a different time on the other side of town. Consider a train timetable which says something like:

08:30	Toronto
08:45	Oakville
09:05	Hamilton

And that's why the railways were among the first to use standard times which were standardized over a larger area than a city. In the late 1840s, the railways in Great Britain adopted "Greenwich Mean Time" (GMT) for all of Great Britain -- all clocks in Great Britain were to be set to the same time, which was the mean solar time over the astronomical observatory in the town of Greenwich, England. GMT was adopted for all other purposes in most of Great Britain gradually over the following decades.

But what to do for a country like Canada? If we were to adopt a single standard time for all of Canada, sunrise would be at a very strange time of day on one side of the country or other. It's too wide, east to west; solar time differs by several hours across the country.

In 1884 a conference was held in the USA called the "Meridian Conference" which established a world-wide system of "time zones", with GMT as a basis. Take one twenty-fourth of the circumference of the Earth, centred around Greenwich; this is the GMT time zone. The next twenty-fourth of the Earth to the east is "+1", and then "+2", and so on. Alternatively, heading west, the next twenty-fourth of the Earth's circumference west of Greenwich is "-1", then "-2", and so on. Eastern Standard Time is time zone "-5" -- the time here is five hours less than GMT, so that when it's noon in Greenwich, it's 07:00 here; or when it's noon here, it's 17:00 in Greenwich (5:00 PM). And this time is standard not just for Toronto, but also the same in Oakville, and Hamilton, and in fact most of Ontario, plus just about all of Québec, and more.

This also means that other time zones are offset from us by a simpler formula. For example, the time in Vancouver is three hours earlier than the time in Toronto... exactly. This is easier than adopting the true mean solar time, which would be offset by an amount which was not an integer number of hours. (I calculate about 174.9 minutes.)

In class we looked at the map you can find at http://www.physicalgeography.net/fundamentals/images/world_time2.gif
You see that the time zones are not on the exact theoretical lines set out by the Meridian Conference; in fact they're awfully wiggly. If a country is mostly in one time zone, but a bit of it crosses the relevant line of longitude, they're going to declare themselves to be all in the one major time zone. It's much easier that way. Similarly, countries with a lot of interchange with each other may find it easier to be in the same time zone where feasible. Also of particular note is China, which is all one time zone even though it's almost as wide (in fraction of the Earth's circumference) as Canada. They've made the decision that being in the same time zone is worth slightly odd sunrise and sunset times at the west and east ends of the country.

The modern name for GMT is "UTC", which stands for "Universal Coordinated Time".

If we think that it would be bizarre to have standard time in Hamilton be two minutes off from standard time in Toronto because of trains, it's all the more important to synchronize time in the two cities when you can be talking between them live on the telephone, or over the internet. It's quite reasonable to phone someone in Hamilton and suggest meeting somewhere which you can both easily get to by train or automobile; or you could schedule a phone call or internet-chat at a particular time with someone living in Hamilton. It would be quite complicated if the two cities had different standard times.

The faster you go, the more this matters. Apart from telecommunications, most people's experience these days with time zones is with airplane travel. The greater speed makes it more of an issue. You rarely have to re-set your watches after an automobile trip, but frequently have to re-set your watches after an airplane trip.

... mentioned Daylight Saving Time here ...

Modern science cares a lot more about precise time, so much so that we can observe that the Earth's rotation is not constant; it varies by small fractions of a second from year to year. But this sort of change to the meaning of a second is not tolerable these days. If a particular phenomenon in atomic physics happens 1,000,000,000 times a second, we don't want to revise all the textbooks next year and say that due to a slight decrease in the Earth's rotational speed (hence lengthening of the day), this phenomenon now happens 1,000,000,003 times a second. This would get ridiculous quite quickly.

The alternative, though, may seem even more bizarre. What has happened is that the second, formerly defined as an 86400th of a day (24 * 3600 == 86400), is now defined as how long it takes an energized Cesium 133 atom to oscillate 9,192,631,770 times -- a scientific definition which, theoretically speaking, has nothing to do with Planet Earth! This amount of time is a 86400th of some days; very very close to an 86400th of others. Over an entire year, the difference is only a fraction of a second.

But over the years these fractional seconds accumulate. Therefore, just as we need inserted days in the calendar to keep the day and year in sync, we now need inserted seconds in the day to keep the day and hour/minute/second in sync. These are called "leap seconds" and they are announced by an organization entitled IERS, "International Earth Rotation Society", based on astronomical observation. The leap second announcements are available on the web at http://www.iers.org/nn_11214/IERS/EN/DataProducts/EarthOrientationData/__Function/generischeTabelle__ID16.html (click on the "text file" button to see the file in your web browser).
(Most of them are boring, because they just say that there is no leap second that year! But the most recent one (C43) is more exciting.)