6

Since it's just come out that Earth is apparently spinning about a milliscecond and a half faster.. I see a bunch of "articles" claiming this will.. somehow.. break atomic clocks and make GPS useless.. (Google "earth spinning faster break atomic clocks" and go to crazy-town..)

Umm, how and what exactly do they think will happen? A couple things I saw (I read two articles, then quit trying to play whack-a-mole to find one that might actually explain the basis of this claim) did admit that we already have to correct for relativity.. And we already have positive leap seconds added every so often, and my understanding is that various ephemeris data has to be updated every so often.. So would this be be really any different?

I can see that if the speed kept increasing by something like.. say two or three seconds per day, or the Earth started wobbling around that could cause problems.. But we're not talking about that (at least until next year)

Qmechanic
  • 201,751
  • 1
    The organisation which monitors the Earth's rotation rate & orientation: https://www.iers.org/IERS/EN/Science/EarthRotation/EarthRotation.html – PM 2Ring Aug 02 '22 at 22:27
  • 1
    The IETF leap seconds list (human & machine readable) https://raw.githubusercontent.com/eggert/tz/main/leap-seconds.list & https://www.ietf.org/timezones/data/leap-seconds.list (the 1st URL is current). – PM 2Ring Aug 02 '22 at 23:07
  • 3
    "Umm, how and what exactly do they think will happen." Yeah, good question. You need to take pretty much everything you read on the Internet with a grain of salt. There's a lot of click-bait baloney out there. – hft Aug 02 '22 at 23:16
  • The question was initially closed. I voted to be reopened (for new answers, if needed). This website could also be used to obtain scientifical explanations on how and why some information spread in the media is wrong. – DanielC Aug 02 '22 at 23:17
  • 5
    Which bunch of articles? – Qmechanic Aug 02 '22 at 23:20
  • 2
    You get better results if you Google "earth spinning faster" instead of "earth spinning faster break atomic clocks". It gets even better if you filter out news sites, but keep scientific sites. That doesn't leave much in this case. But read When a Day is Not a Day from NASA. – mmesser314 Aug 03 '22 at 01:46
  • 4
    The internet is insane. Stop being baffled by it. The experimental evidence is overwhelming. – Jon Custer Aug 03 '22 at 02:08
  • By calling it "insanity" you already disregard the possibility that they are correct. – Ivo Aug 03 '22 at 07:21
  • " everything you read on the Internet with a grain of salt." There's a

    – Jim Witte Aug 04 '22 at 01:52
  • 1
    https://arstechnica.com/science/2022/08/record-short-days-could-speed-up-debate-on-leap-seconds/ – John Doty Aug 04 '22 at 18:41

4 Answers4

20

None of the articles (e.g. this one) I came across say anything about breaking atomic clocks. The problem is simply a logistical one - if we have to implement a negative leap-second, then we will be doing something which most of the software which implements ordinary leap-seconds has not been explicitly designed to do. And if you've ever written any large-scale computer code, you are probably aware that a tiny change can wreak massive havoc if that change hasn't been thoroughly tested.

J. Murray
  • 69,036
  • 1
    Perhaps the IERS could use this as an opportunity to get rid of the !@#$% leap seconds. – John Doty Aug 02 '22 at 22:17
  • 2
    @JohnDoty that would require agreement on one of the alternatives. That's unlikely. – BowlOfRed Aug 02 '22 at 22:27
  • 2
    @JohnDoty You may enjoy perusing https://www.ucolick.org/~sla/leapsecs/ – PM 2Ring Aug 02 '22 at 22:30
  • 2
    From https://www.ucolick.org/~sla/leapsecs/right+gps.html The POSIX standard requires that the value of time_t have a simple relationship to date and time where each day has 86400 seconds. Such a requirement demands that POSIX systems not count leap seconds because they produce days which have 86401 seconds.

    The POSIX standard also specifies that the time and date values be UTC. UTC has had leap seconds, so the combination of the two requirements is inconsistent with the broadcast time signals that are used to set system clocks.

    – PM 2Ring Aug 02 '22 at 22:58
  • If you are interested in IERS you may be interested in the IGS products obtained from GPS laser ranging, including the clock products https://igs.org/products/ – UVphoton Aug 02 '22 at 23:48
  • @BowlOfRed Not so hard. Just stop issuing them. – John Doty Aug 03 '22 at 00:32
  • @PM2Ring Yep. That's one of many problems. – John Doty Aug 03 '22 at 00:35
  • 1
    I work for a company that sells atomic and other accurate clocks. Even the implementation of positive leap seconds in some NTP clients is patchy and I have seen whole networks fall over when, e.g. their routers crash on a pending leap second indication. I am scared for the day a negative leap second is introduced. – Darren Aug 03 '22 at 06:49
  • If time_t counted leap seconds, not only would the time_t % 86400 == 0 at midnight relationship break, there would be no deterministic relationship between time_t and future moments in the calendar, since this relationship would change each time a leap second was added, and they are added at short notice. – Rodney Aug 03 '22 at 09:32
  • @Darren: Meh, just smear your leap seconds and then you're basically good. Over a 24-hour period, you'll never experience a discrepancy with the outside world larger than 0.5 seconds, which is only slightly worse than typical network latency, so most systems should tolerate it just fine. – Kevin Aug 03 '22 at 15:34
  • @Kevin unfortunately, applications that have to be accurate to the nano second scale will notice it. – Darren Aug 03 '22 at 15:54
  • 1
    @Darren: Sure, scientific applications will care, but IMHO they should be using TAI or GPS time instead of UTC in the first place. – Kevin Aug 03 '22 at 15:55
  • @PM2Ring POSIX being poorly designed to the point it can't deal with the complexities of proper timekeeping isn't an argument against proper timekeeping – ScottishTapWater Aug 04 '22 at 09:27
  • 1
    @ScottishTapWater I wasn't implying that. I was merely pointing out that POSIX is a problem. However, it's not easy to get everyone to agree on what "proper timekeeping" even means, as I said in https://physics.stackexchange.com/questions/721236/earth-spinning-faster-will-break-gps-and-atomic-clocks-umm-how?noredirect=1#comment1613332_721239 – PM 2Ring Aug 04 '22 at 10:15
  • @PM2Ring Fair one, I misunderstood the intent of your comment... My bad – ScottishTapWater Aug 04 '22 at 15:47
  • @Kevin Yes, use TAI for your data taking, but then you run into the practical problem that computers and their networks use UTC. So, in your infrastructure, you have interfaces that don't stay in sync. – John Doty Aug 04 '22 at 18:50
  • @JohnDoty: There are two options: You care about being in sync with the internet, or you don't. If you don't, then fine, you just run the whole system on TAI and ignore internet time altogether (maybe you do some sort of conversion to UTC at a later time, after you've gathered your data and are getting ready to analyze it). If you do care about the internet, then you have to contend with tens or hundreds of milliseconds of network lag anyway, and an extra 500 milliseconds is not really that bad in the grand scheme of things. – Kevin Aug 04 '22 at 21:21
  • @Kevin But in a big, complicated endeavor like a space mission, you can't choose one or the other. Some parts of your operation use UTC, some parts something else. – John Doty Aug 04 '22 at 21:32
  • @JohnDoty: In a space mission, you should not be using commercial software in the first place (at least, for anything that actually matters, in the sense that the mission will fail if it messes up), and so you can meticulously audit every line of code that deals with time to make sure it's leap-second safe (or just, y'know, don't plan a space mission at the same time as a leap second in the first place). This is not really an option for the broader software industry, which is the purpose for which the leap smear was designed, because it's too expensive for not enough value. – Kevin Aug 05 '22 at 03:22
  • @Kevin In reality, space missions employ large amounts of commercial software. Most of a space mission's infrastructure is on the ground, and very few ground computers use custom operating systems. Even in space, these days most spacecraft use a POSIX operating system (often Linux), and the POSIX standard does not deal with leap seconds gracefully. Plan a space mission to avoid leap seconds? Missions take years to plan, and often last decades. Leap seconds are announced a few months in advance. – John Doty Aug 05 '22 at 14:37
6

I would like to see a real expert discuss, but I think from what I can tell that these clickbait articles started from Forbes magazine

If Earth spins faster then it gets to the same position a little earlier. A half-a-millisecond equates to 10-inches or 26 centimetres at the equator. In short, GPS satellites—which already have to be corrected for the effect of Einstein’s general relativity theory (the curve of space and time)—are quickly going to become useless.

However my understanding is that gps satellite orbit data and clock data are constantly being updated and corrected for and that the IGS has a series of products that update regularly with these orbits and clock corrections.

IGS is the International Global Navigation Satellite Service and has a network of about 500 ground stations that is gathering the orbit and clock data and probably some ionosphere data as well.

The reason the orbit data is so important is that orbits constantly change and without updates would be kilometers out of position in a relatively short time, while for GPS, I think satellite positions are kept within a few centimeters and clock times are kept to the order of a few 10s of nanoseconds for real time positioning applications. The gps position error on the ground is a different much larger number.

So while you can certainly see the difference in clock due to general relativity, and the earth spun faster, I don’t think it translates to any unsolvable or dangerous issues.

UVphoton
  • 3,283
  • 4
    FWIW, GPS time doesn't use leap seconds. https://gssc.esa.int/navipedia/index.php/Time_References_in_GNSS – PM 2Ring Aug 02 '22 at 23:31
  • 1
    That’s my point. IGS and the ground stations produce a wide variety of clock and position products to very high accuracy using tools like laser and radar ranging. https://igs.org/products/ – UVphoton Aug 02 '22 at 23:37
  • If we could do everything in GPS time, there would be no problem. However, civil time is UTC, which means that time used by computer systems isn't in sync with GPS time. – John Doty Aug 03 '22 at 01:39
  • 2
    @JohnDoty It's not that simple. From https://www.ucolick.org/~sla/leapsecs/picktwo.html "The proposals to abandon leap seconds in UTC have not included text which describes the effect that would have on disconnecting the calendar from the rotation of the earth". Modern precision timekeeping involves multiple groups of people, and these groups don't necessarily understand or communicate clearly with each other. These include various engineering & science disciplines, and civil, military and even religious authorities. Reaching some form of agreement & compromise can be difficult... – PM 2Ring Aug 03 '22 at 02:22
  • 1
    The Forbes article is rather silly. As this answer says, the orbits of the GPS satellites are constantly changing, affected by eg, inhomogeneities in the earth, and changes in its rotation speed (things which IERS exists to keep track of), and a deviation of a few ms here or there is well within the range the overall system is designed to cope with. There are multiple timescales – TAI, GPS, UT(0/1/2), UTC, NTP/unixtime, ... – each the right thing for a different purpose, and the interrelationships between which are... funky. – Norman Gray Aug 03 '22 at 15:07
1

Atomic clocks will keep running forever at a speed of exactly one second per second. Earth rotation might slow down from say 86,400 seconds a day to 86,399.999 seconds a day. GPS will perfectly happily find the exact point in space and time where you are, but for example your home might not be where it is expected to be found anymore. That's because your home is rotating at a speed of (I think) 1,700 km per hour around earth.

So GPS has to take into account the earth rotation, and when it has found your absolute 3D coordinates in space, it has to find out where earth exactly is, how far it has rotated through the day, and from that it calculates where it's 3D coordinates are on a map of the earth.

(GPS systems do some tricks to get better results once they have your position on a map of the earth. If GPS finds only three satellites, it cannot find your 3D location and the time simultaneously. It will then goes that most likely you are somewhere on the ground. Say if you are roughly in an are 2,000 feet above sea level, then it guesses that your 3D location is 2,000 feet above sea level - if you are on top of a high building, that guess may be wrong. And if you are driving, it guesses that you are on some street. If there is no street on the map where you are, but there is a street 10 feet to the east, it may guess that you are ten feet to the east).

In practice, you put GPS receivers say in 10 places on the earth. Then you compare what the GPS receivers think where they are, with the unchanged actual location of the receivers (unchanged unless there is a huge earthquake). With this information, you calculate how far an uncorrected GPS receiver would be wrong, and then you fix its calculation. That is actually one way how you measure the speed of earths rotation.

  • The answer could be improved by removing the paragraph in parenthesis about what GPS receivers do. That only distracts from the main point that the position and timing of the satellites is regularly synced to ground station data. – fishinear Aug 03 '22 at 12:40
  • Yes, but how does the ground station know where it is? It knows where it is on the earth, but it doesn't know where the satellite thinks it is, unless you ask the satellite. – gnasher729 Aug 03 '22 at 17:15
  • I meant removing the paragraph that you have put in parenthesis (currently the third paragraph), because it only adds less relevant information that distracts from how the GPS system works. – fishinear Aug 03 '22 at 17:23
  • As the Earth's rotation rate decreases the number of SI seconds in a mean solar day increases. Over the last few decades the mean solar day length has been ~86400.002 SI seconds. The SI second was defined in terms of Cs-133 in 1967, but the duration adopted for it matched the then-current ephemeris second, which is equal to the mean solar second from ~1820. I have some relevant info & links at the end of this answer. – PM 2Ring Aug 03 '22 at 18:08
1

Atomic clocks

Atomic clocks do not deal with leap seconds and the length of a day. All atomic clocks internally just keep a counter that indicates the number of seconds (or rather picoseonds) that have past since some universally agreed point in time in the past. Mid-night on January 1st, 1970 in Greenwich is such an universally used common time point in the past. Time synchronisation between atomic clocks is also based on such a standard.

Computers

EDIT: the below description for computers is correct for leap days, time zones, and such. But apparently I assumed too hastily that it applied to leap seconds as well. Leap seconds are actually incorporated into the running count on some computers. Thanks for Ilmari for pointing that out.

All clocks in computers also just keep a running counter of number of seconds since a fixed time point. They then synchronise that counter over the Network Time Protocol with the internet to keep the clock in sync with other clocks.

It is only when translating those counters to human readable times, that leap seconds (and leap days, time zones, and summer/winter time) come into play. We humans like to have the hours, minutes, and seconds that we use to be correlated to the spinning of the Earth, such that the sun is always directly above at 12:00. For that reasons, computers translate the internal clock counter to a human readable time, according to mapping tables in the operating system. Those tables also take care of all the different time zones in the present and both the past and future.

Those translation tables need to adapted when the Earth spins slower. But that is not a dramatic change. If it is agreed in advance when a second is going to be skipped instead of added (a negative leap second), then those tables can just be adjusted in advance without major consequences.

Sattelite navigation systems

For satellite navigation systems, they are not affected by this mapping at all. The clocks in navigation satellite don’t deal with human times either. Only two things are important for the satellites:

  1. that each satellite knows precisely above which geographic location it is located at any point in time.
  2. that the clocks in the satellites are synchronised with each other, such that it is guaranteed that they all show the same time from an earth surface point of view.

Due to general relativity, clocks on satellites run faster than on Earth, and need to be compensated for that. Satellite orbits can also not be predicted with perfect accuracy. For both these effects, there are already adjustment mechanisms in place. Both clock speed and the orbit information of the satellites are regularly synced with ground stations in order to make sure the information remains accurate. A milli-second more or less is not going to upset those adjustment mechanisms.

fishinear
  • 884
  • 1
    Alas, while what you describe (i.e. effectively keeping the internal clock synchronized with TAI) would indeed be a sensible way for computers to handle leap seconds, it's not the way most computers currently do it. In practice, most computers today use some variation of "unix time" for their internal timekeeping, typically synced to (current) UTC but calculated with the (incorrect) assumption that every UTC day is 86400 seconds long. See the linked Wikipedia article for more details. – Ilmari Karonen Aug 03 '22 at 14:42
  • 1
    … Part of the problem is that the contradictory assumptions — i.e. that 1) unix time increments by one every SI second, 2) the offset between unix time and UTC is constant, and 3) every UTC day is 86400 unix seconds long — are not just programmed into the OS, but also into literally countless programs, libraries and programming languages (and sometimes even into standards defining those languages). Also, unix timestamps are frequently exchanged between different programs on different computers, so any changes to their interpretation relative to UTC would cause interoperability issues. – Ilmari Karonen Aug 03 '22 at 15:16
  • … So, in practice, it's assumption #1 above that gets broken, not #2 or #3. (Ps. See also this Q&A thread on Stack Overflow.) – Ilmari Karonen Aug 03 '22 at 15:16