Results 1 to 21 of 21

Thread: I don't get leap seconds!

  1. #1
    Join Date
    Nov 2004
    Posts
    5,564

    I don't get leap seconds!

    Why does the slowing of the Earth's rotation by 2 milliseconds every dozen decades necessitate the adding of an entire second every 2-4 years?
    "Occam" is the name of the alien race that will enslave us all eventually. And they've got razors for hands. I don't know if that's true but it seems like the simplest answer."

    Stephen Colbert.

  2. #2
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,904
    Quote Originally Posted by parallaxicality View Post
    Why does the slowing of the Earth's rotation by 2 milliseconds every dozen decades necessitate the adding of an entire second every 2-4 years?
    If our master clock is running at the former rate, the Earth will get behind by about 2 milliseconds each day, or a full second in about a year and a half. In recent decades the rotation has speeded up because of some sort of mass shift, making the need for leap seconds less frequent. Over the long haul the tidal interaction will resume slowing it down, making the need for leap seconds more and more frequent.

    A friend of mine told me her husband, an astronomer in the timekeeping department of the U.S. Naval Observatory, would prefer to see leap seconds go away. In principle we the people of Planet Earth could by law enact a one-hour adjustment whenever the Earth gets more than some 30 minutes out of sync, analogous to enacting periodic switches back and forth between standard and daylight time. I recently saw a remark about the possibility of timekeeping-related legal problems with far less lag, but in my opinion they were overthinking heaven only knows what.

  3. #3
    Join Date
    Nov 2004
    Posts
    5,564
    If our master clock is running at the former rate, the Earth will get behind by about 2 milliseconds each day, or a full second in about a year and a half.
    Sorry to beat this into the ground but why?
    "Occam" is the name of the alien race that will enslave us all eventually. And they've got razors for hands. I don't know if that's true but it seems like the simplest answer."

    Stephen Colbert.

  4. #4
    Join Date
    May 2003
    Posts
    6,113
    Quote Originally Posted by parallaxicality View Post
    Sorry to beat this into the ground but why?
    If your clock is set to have 86,400 seconds per day, but each actual rotation takes 86,400.002 seconds, you get a little more out of sync with the rotation every day. After two days you'll be 0.004 seconds off, and so on. That accumulates unless you re-synchronize occasionally.
    Conserve energy. Commute with the Hamiltonian.

  5. #5
    Join Date
    Sep 2005
    Location
    Metrowest, Boston
    Posts
    4,789
    parallaxicallity. The why is angular momentum. Mass at a radius from a point of rotation. The Earth rotates and has a lot of angular momentum. Some of it transfers to the moon due to tidal effects, causing the moon to slowly move to a more distant orbit...cm/ year. And, there's the melting ice issue. If all the ice was on mountains near the equator, when they melt and transfer water to the oceans, the Earth should speed up, like an ice skater drawing in her/ his arms while spinning, by Conservation of angular momentum.
    But, the ice is at the polar regions, rotates slowly, and much is near sea level. So, the "new" water is mostly in the Pacific and Atlantic oceans near the equator. That speeded up water subtracts it's angular momentum out of the already spinning blue planet Earth, making it slow down.
    pete

  6. #6
    Join Date
    Jul 2005
    Posts
    18,484
    Quote Originally Posted by parallaxicality View Post
    Why does the slowing of the Earth's rotation by 2 milliseconds every dozen decades necessitate the adding of an entire second every 2-4 years?
    It doesn't. It's not actually the continuing gradual rate of slowing that's the problem, it's the fact that the solar day, right now, is 86400.002 seconds long. Even if the rate of slowing was zero, we'd need the leap seconds to compensate for the existing mismatch between our clocks and the mean solar day. All the gradual continuing slowing does is modulate how often, in the long term, we would need to add leap seconds.

    Grant Hutchison

  7. #7
    Join Date
    Nov 2004
    Posts
    5,564
    Ah. OK I get it.
    "Occam" is the name of the alien race that will enslave us all eventually. And they've got razors for hands. I don't know if that's true but it seems like the simplest answer."

    Stephen Colbert.

  8. #8
    Join Date
    Jul 2018
    Posts
    146
    Quote Originally Posted by grant hutchison View Post
    It doesn't. It's not actually the continuing gradual rate of slowing that's the problem, it's the fact that the solar day, right now, is 86400.002 seconds long. Even if the rate of slowing was zero, we'd need the leap seconds to compensate for the existing mismatch between our clocks and the mean solar day. All the gradual continuing slowing does is modulate how often, in the long term, we would need to add leap seconds.

    Grant Hutchison
    Wow, thats awesome. That makes complete sense now, I too was a bit confused by the apparent mismatch parallax pointed out. I will file this under "things I didn't even know I wanted to know, but it turns out I do". Thanks!

  9. #9
    Join Date
    Jul 2005
    Posts
    18,484
    Yeah, I think the story of the leap second is sometimes summarized down to such a condensed state that it's difficult to understand.
    We do have leap seconds because the Earth's rotation rate is slowing, but the reason we have so many is because of the way the duration of the second was defined and then redefined over a long period of time - at first astronomically, and then switched to a definition in terms of physics. This meant that the SI second (when defined physically half a century ago) was based on the length of the day as measured by astronomers a century previously. Hence the 2ms mismatch we now enjoy, which brings the leap seconds around fairly frequently.
    The other problem is that the average slowing, of 1.4-1.7ms/dy/century, obscures a relatively large amount of year-to-year variation, which is of magnitude comparable to a second - so the leap-second interval can come more or less frequently than what you'd predict from the simple 2ms average daily mismatch. There's even the potential to need a negative leap second, but we haven't reached that threshold so far.

    Grant Hutchison
    Last edited by grant hutchison; 2019-Dec-14 at 02:23 PM.

  10. #10
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,904
    Quote Originally Posted by Hornblower View Post
    If our master clock is running at the former rate, the Earth will get behind by about 2 milliseconds each day, or a full second in about a year and a half. In recent decades the rotation has speeded up because of some sort of mass shift, making the need for leap seconds less frequent. Over the long haul the tidal interaction will resume slowing it down, making the need for leap seconds more and more frequent.

    A friend of mine told me her husband, an astronomer in the timekeeping department of the U.S. Naval Observatory, would prefer to see leap seconds go away. In principle we the people of Planet Earth could by law enact a one-hour adjustment whenever the Earth gets more than some 30 minutes out of sync, analogous to enacting periodic switches back and forth between standard and daylight time. I recently saw a remark about the possibility of timekeeping-related legal problems with far less lag, but in my opinion they were overthinking heaven only knows what.
    Another possibility in the long term would be to change the value definition of the SI second to better fit Earth's spin rate over the next few centuries. However, that would mean adjusting a lot of fundamental constants and would only be a temporary fix. I think it is better to stick with the fundamental constant values we have now and occasionally tweak our civil timekeeping. The varying spin rate of our less than perfect planet is really of no consequence in the larger world of cosmological physics.

  11. #11
    Join Date
    Feb 2009
    Posts
    2,246
    Quote Originally Posted by Hornblower View Post
    Another possibility in the long term would be to change the value definition of the SI second to better fit Earth's spin rate over the next few centuries. However, that would mean adjusting a lot of fundamental constants and would only be a temporary fix.
    Yet that was what was done until 1972 - stretching the second to correct for the slowing rotation of Earth. Leap seconds were introduced from 1972 by the decision to freeze the length of the second, at about the average over the period 1750-1950.

    Leap seconds are only allowed 2 times a year. When will this run out and require a different correction?

  12. #12
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,904
    Leap seconds will be needed more than twice a year when the daily lag exceeds about 5.5 milliseconds. If the recent anomalous speedup runs its course and the long term slowdown resumes, that will happen in a little over 3 centuries.

  13. #13
    Join Date
    Jul 2005
    Posts
    18,484
    This paper may be of interest: "The Leap Second: Its History And Possible Future" (Nelson et al., 2001)

    Grant Hutchison

  14. #14
    Join Date
    Aug 2019
    Posts
    395
    I have never gotten the whole leap second and leap year thing also.

  15. #15
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,904
    Quote Originally Posted by Sinbad View Post
    I have never gotten the whole leap second and leap year thing also.
    If you still don't understand it, here is a rough explanation. Let's start with the leap year. The Earth's spin rate is such that it needs 365 days and about 6 hours to make a complete circuit of its orbit, or 1,461 days for four circuits. By declaring one calendar year out of every four to be 366 days instead of 365, we get the necessary 1,461 days over a period of four years, and the calendar stays nearly in sync with the seasons.

    The leap second is according to the same principle. The mean solar day is not an exact whole number of standard atomic-time seconds.

  16. #16
    Join Date
    Aug 2005
    Location
    NEOTP Atlanta, GA
    Posts
    2,957
    This video from Scott Manley may help to explain why leap seconds (and days) exist.

    https://youtu.be/2mUTpB8CVpo


    Sent from my iPad using Tapatalk

  17. #17
    Join Date
    Aug 2019
    Posts
    395
    No I understand why do it. I just don’t get the importance like daylight savings time

  18. #18
    Join Date
    Nov 2004
    Posts
    5,564
    Daylight savings time resets itself every year. It's completely optional. Leap seconds are like leap years; if we don't add them then our clocks become ever more out of sync.

    There are two different kinds of time; atomic time, which is just an endless count of caesium pulses tick-tocking unstoppably into the future, and Universal Coordinated Time, which is atomic time but adjusted for all those traditional things like "days" and "years".

    Atomic time will never be reset, so UTC has to constantly readjust to keep up with it.
    "Occam" is the name of the alien race that will enslave us all eventually. And they've got razors for hands. I don't know if that's true but it seems like the simplest answer."

    Stephen Colbert.

  19. #19
    Join Date
    Aug 2019
    Posts
    395
    That makes more sense now

  20. #20
    Join Date
    Jun 2003
    Location
    Central Florida.
    Posts
    5,936
    I did some 'round-the-clock near-realtime tracking of Titans and Shuttles for a NASA contractor. One support ran past Jun 30, and on July 1 we were told to correct for a leap second that had been implemented. The program we were using didn't haven an explicit correction method, so we simply moved the entered liftoff time back by one second, which made all the plus-time computations come out as expected.* It probably wasn't essential, but I found it interesting.

    *In another experimental operation, we were given data in Julian Seconds rather than the Modified Julian Seconds expected. Rather than translate the value in the input file, or correct it each time, we also set back the liftoff time by enough days to effectively strip off the most significant digits. However, in this case the entered liftoff time was some date in 1853 or thereabout! Looked funny, but it got the job done.

  21. #21
    Join Date
    Aug 2019
    Posts
    395
    Makes sense

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •