Exploring the Mystery of Earth-Centered Time Reference

In summary: They flew four cesium clocks completely around the world, once in the easterly direction and once in the westerly direction, and compared the results in both cases with surface clocks.
  • #1
bwana
82
2
Haefele and Keating flew an atomic clock from the west coast to the east coast in 1971 and then they flew it back. The east going clock lost 59 +- 10 ns and the west going clock gained 273 +- 7 ns. What I do not understand is why they used the center of the Earth as a reference frame, and not the surface-after all that's where the surface clock is located. They said:

"The problem encountered with measuring the difference between a surface clock and one on an aircraft is that neither location is really an inertial frame. If we take the center of the Earth as an approximation to an inertial frame, then we can compute the difference between a surface clock and the aircraft clock. "

The need to establish a third frame of reference-the center of the earth-is confusing. Why is neither the surface clock's nor the airborne clock's frame of reference valid? Because the Earth's surface is moving? But that's a major point of the einstein equivalence principle-that moving frames of reference have the same physics.

I understand the equation for the time difference calculation:
Ta-Ts=To[(2Rwv+v^2)/2c^2]
where Ta is airborne time, Ts is surface time, To is center of the Earth time
R is Earth radius, w is Earth angular velocity, v is plane velocity, c is light speed
And then they go on to use the following approximation:
Ta-Ts=-Ts[(2Rwv+v^2)/2c^2]
The difference being they use surface time instead of center of the Earth time. They state:
Note that the "earth center" time has been replaced by the surface time in this expression. This is a valid approximation in this case since the time difference is many orders of magnitude smaller than the time itself, and this allows us to model the difference between two measurable times.
How can they conclude by using Ts when they started the whole analysis saying that surface time was not a proper inertial frame?

And another confusing aspect of using the center of the Earth as the reference frame is if you change the flight plan a little bit. Suppose the plane flies along a line of longitutude instead of latitude? Since v is orthogonal to Earth angular velocity, is it measured as 0? A flight to south pole is the same as a flight to north pole in terms of time dilation? I wonder if that expt has been done?
 
Physics news on Phys.org
  • #2
bwana said:
Haefele and Keating flew an atomic clock from the west coast to the east coast in 1971 and then they flew it back.

No, they flew four cesium clocks completely around the world, once in the easterly direction and once in the westerly direction, and compared the results in both cases with surface clocks.

bwana said:
Why is neither the surface clock's nor the airborne clock's frame of reference valid? Because the Earth's surface is moving?

The quote you mentioned didn't say those frames were not "valid", it said they were not inertial. The reason they are not inertial frames is two-fold. First, they are obviously not moving inertially. (They are going around in a circle.) Second, they are located at different heights within a gravitational field.

bwana said:
But that's a major point of the einstein equivalence principle-that moving frames of reference have the same physics.

You're mis-informed. The theory of relativity does not assert that the descriptions of phenomena are the same regardless of the system of coordinates in terms of which they are described, let alone that they are the same in terms of non-inertial coordinate systems as they are in terms of inertial coordinate systems. When people say the "physics is the same" for all (diffeomorphically equivalent) systems of coordinates in general relativity they mean that the same equations of physics, i.e., the field equations, are equally applicable to any such coordinates. This is true, but to use simplified equations that are adapted to specialized coordinate systems (e.g., inertial coordinates) we must obviously use the specialized coordinate systems. Otherwise the simplified equations are not applicable.

bwana said:
How can they conclude by using Ts when they started the whole analysis saying that surface time was not a proper inertial frame?

This is just simple algebra. The quantity "To" can be written as Ts(1+q), where q is an extremely tiny dimensionless number (with suitable sign). Also the factor in the square brackets is an extremely tiny dimensionless number, let's call it p. So the time difference is

Ta-Ts = -To[p] = -Ts(1+q)p = -Ts[p] - Ts(qp)

Obviously the second term in that last expression (containing the factor of both q and p) is tiny compared to the first term, so it's fine to represent Ta-Ts as -Ts[p], provided only that we have an accurate evaluation of the factor p, which of course requires us to correctly account for the difference between To and Ts (and Ta).

bwana said:
Suppose the plane flies along a line of longitutude instead of latitude? Since v is orthogonal to Earth angular velocity, is it measured as 0?

No, v would not be measured as zero. Being orthogonal to the tangential velocity of the Earth's surface does not imply that it is zero.

bwana said:
A flight to south pole is the same as a flight to north pole in terms of time dilation?

Yes, a flight from the north pole to the south pole (along a constant "longitude" line mapped on the Earth's surface) would be symmetrical to a flight from south pole to north pole - at least if you neglect the effect of the Earth's orbital motion around the Sun.
 
Last edited:
  • #3
Tnx samshorn for ur prompt reply.
1)yes i was inaccurate in describing their procedure. They did fly around the world. I also assumed that the surface clock was at the same latitude as the aircraft which flew circumferentially around the planet at a constant latitude, but I might be wrong there as well.
2) I am only addressing the kinematic contribution to SR. The flights in fact were at different altitudes and the gravitational contribution to time dilation was different on the different parts of their journey. I am not questioning that. But I do not understand your statement that the surface clock and the airborne clock are in reference frames which " are obviously not moving inertially." What is an example of two reference frames that ARE moving inertially? When Einstein described the light clock on a train and a passenger watching it while it passed him at the station, were the passenger and the train considered to be in inertial frames?
3) How can we assume that To=Ts(1+q) where q is small? To is the time at the center of the earth. There is no gravity there so time dilation should be greatest there. I would guess that To is radically different than Ts.Consider instead putting the surface clock at the north pole. Ignoring for the moment that the Earth is an oblate spheroid and the gravitational effects are slightly different at the poles, consider the kinematic component. Angular velocity of the surface clock at the pole is 0. In this case, I understand how there can be an asymmetry of SR because the aircraft's velocity is relative to ground speed. The total aircraft velocity is therefore radically different depending on its direction.

As a corollary, I am still pondering the possibility that there is time dilation at the equator due to rotational velocity. So people age faster at the poles?
 
  • #4
bwana said:
As a corollary, I am still pondering the possibility that there is time dilation at the equator due to rotational velocity. So people age faster at the poles compared with people at the equator?

I added a few words to your quote, because we should never say "faster" without being clear about what we're comparing with. But with that said, and if we simplify the problem by treating the Earth as a perfect sphere (this is a plenty good enough approximation here), the answer is "yes".

You might try calculating just how much the difference is... you can find the time dilation formula all over the Internet and you can calculate the relative velocity between poles and equator because the circumference of the Earth is about 40000 kilometers and it rotates on its axis once every 24 hours.
 
  • #5
bwana said:
I do not understand your statement that the surface clock and the airborne clock are in reference frames which " are obviously not moving inertially." What is an example of two reference frames that ARE moving inertially?

Inertial motion is defined as (locally) unaccelerated motion. Hence anything that is unaccelerated is an example of something that is moving inertially. This is the same as in Newtonian mechanics, so your problem has nothing in particular to do with special or general relativity. I think it's hard for someone who doesn't have a working knowledge of Newtonian mechanics (or quantitative physical reasoning in general) to understand relativity.

bwana said:
When Einstein described the light clock on a train and a passenger watching it while it passed him at the station, were the passenger and the train considered to be in inertial frames?

If you mean "were they considered to be at rest in inertial frames", the answer is a qualified Yes, to the level of precision necessary for those scenarios, i.e., over a short and virtually flat span on the Earth's surface, and over an extremely short duration of time (the nanoseconds it takes for light to traverse the train car), the deviations of the objects from inertial motion are negligible. In contrast, if the train continued all the way around the globe in a circle, and if we consider the station over a period of 24 hours circling the center of the earth, the objects are obviously deviating very significantly from being at rest in any single inertial coordinate system for the entire time. Regardless, we can always compute the effects of the non-inertial motion. In some circumstances the effects are negligible, and in others they are not. Exactly the same is true in Newtonian mechanics, so your problem is not with relativity, per se, it is with quantitative reasoning in physics generally.

bwana said:
How can we assume that To=Ts(1+q) where q is small? To is the time at the center of the earth. There is no gravity there so time dilation should be greatest there. I would guess that To is radically different than Ts.

Your guess would be radically wrong. Fortunately there's no need to guess. General relativity tells you the difference in time dilation between the surface and center of the Earth, and it is extremely small.

bwana said:
Ignoring for the moment that the Earth is an oblate spheroid and the gravitational effects are slightly different at the poles, consider the kinematic component. I am still pondering the possibility that there is time dilation at the equator due to rotational velocity. So people age faster at the poles?

Well, neglecting differences in gravitational potential, etc., if you just imagine an object at rest (to the necessary precision) in an inertial coordinate system and another object moving around it in a circle, then yes, the object at rest in the inertial frame will accumulate more elapsed time.
 
  • #6
Nugatory said:
if we simplify the problem by treating the Earth as a perfect sphere (this is a plenty good enough approximation here)

Not really, because the Earth's oblateness *does* have a significant effect on clock rates for observers on its surface. In fact, it has just enough effect, due to increased altitude from oblateness, to compensate for the effect of rotational velocity, so clocks everywhere on the Earth's surface go at the same rate. So clocks at the poles do *not* run faster (or slower) than clocks at the equator, if both sets of clocks are at rest with respect to the (rotating) Earth.

Strictly speaking, this is true for an idealized Earth that is a perfect oblate spheroid--the real Earth does deviate from this by small amounts from place to place. But it's still a much better approximation than the spherical one for this question. One reason is that it illustrates an important fact about rotating bodies: the "natural" configuration for such a body is for its surface to be an equipotential surface, i.e., for it to be such that clocks everywhere on the surface go at the same rate. If the surface is not an equipotential surface, there will be stresses within the object that will tend to shift the surface towards being an equipotential surface; i.e., the equipotential surface is the stationary equilibrium state for a rotating object.

I understand that the questioner was trying to ignore the effects of altitude; but I suspect that is because he thought the effects of altitude would be negligible for the poles vs. the equator. They're not.
 
  • #7
PeterDonis said:
Not really, because the Earth's oblateness *does* have a significant effect on clock rates for observers on its surface.

So it does... Next time I'll say "idealization" or "simplification".
 
  • #8
So here is the nut of the problem

Samshorn said:
Inertial motion is defined as (locally) unaccelerated motion. ...
If you mean "were they considered to be at rest in inertial frames", the answer is a qualified Yes, to the level of precision necessary for those scenarios, i.e., over a short and virtually flat span on the Earth's surface, and over an extremely short duration of time (the nanoseconds it takes for light to traverse the train car), the deviations of the objects from inertial motion are negligible...

I appreciate your continued clarifications. So, if I understand this correctly, (and continuing to ignore gravity as a simplification), if I a ringworld spinning at a very high velocity (~ 0.01 c at its equator) and did a similar experiment with a train riding on the inner surface that could reach a speed of 0.000001 c, then I could use a surface clock as an approximation of the inertial rest frame because the difference between the ringworld and the train would be so small?
 
  • #9
bwana said:
So, if I understand this correctly, (and continuing to ignore gravity as a simplification), if I a ringworld spinning at a very high velocity (~ 0.01 c at its equator) and did a similar experiment with a train riding on the inner surface that could reach a speed of 0.000001 c, then I could use a surface clock as an approximation of the inertial rest frame because the difference between the ringworld and the train would be so small?

Whether a point on the surface approximates inertial motion depends (by definition) on the acceleration of that point and the relevant duration under consideration, neither of which you've specified. All you have specified is the speed of the surface, which tells us nothing (by itself) about the deviation from inertialness. You need to specify the acceleration (e.g., by specifying the radius of the surface, from which, given the speed, we could infer the acceleration), along with the relevant time duration for the events under consideration.

Your concluding phrase ("because the difference between the ringworld and the train would be so small") is also puzzling. Again, the deviation from inertialness of any object just depends on its acceleration and the duration. It is not determined by the smallness of the "difference" (whatever kind of difference you have in mind) between two objects - unless you are stipulating that one of those objects is inertial.
 
Last edited:
  • #10
PeterDonis said:
Not really, because the Earth's oblateness *does* have a significant effect on clock rates for observers on its surface. In fact, it has just enough effect, due to increased altitude from oblateness, to compensate for the effect of rotational velocity, so clocks everywhere on the Earth's surface go at the same rate. So clocks at the poles do *not* run faster (or slower) than clocks at the equator, if both sets of clocks are at rest with respect to the (rotating) Earth.

Strictly speaking, this is true for an idealized Earth that is a perfect oblate spheroid--the real Earth does deviate from this by small amounts from place to place. But it's still a much better approximation than the spherical one for this question. One reason is that it illustrates an important fact about rotating bodies: an important fact about rotating bodies: the "natural" configuration for such a body is for its surface to be an equipotential surface, i.e., for it to be such that clocks everywhere on the surface go at the same rate. If the surface is not an equipotential surface, there will be stresses within the object that will tend to shift the surface towards being an equipotential surface; i.e., the equipotential surface is the stationary equilibrium state for a rotating object.
I've always wondered if this was just a coincidence but you're saying it's not. So the same would be true for Mars? Does the fact that the Moon, Mercury and Venus have synchronous rotations disqualify them from the same effect?
PeterDonis said:
I understand that the questioner was trying to ignore the effects of altitude; but I suspect that is because he thought the effects of altitude would be negligible for the poles vs. the equator. They're not.
At least he's in good company. Einstein predicted the same thing at the end of section 4 of his 1905 paper introducing SR:

Thence we conclude that a balance-clock at the equator must go more slowly, by a very small amount, than a precisely similar clock situated at one of the poles under otherwise identical conditions.
 
  • #11
ghwellsjr said:
At least he's in good company. Einstein predicted the same thing at the end of section 4 of his 1905 paper introducing SR

Well, it's a bit unfair to fault Einstein for failing to take into account the effects of general relativity before he discovered it.:biggrin:
 
  • #12
ghwellsjr said:
I've always wondered if this was just a coincidence but you're saying it's not. So the same would be true for Mars?

Yes, at least to a first approximation. As I said, the Earth is not exactly an oblate spheroid, because its substance, being mostly solid (part of the core is liquid), can resist the stresses set up by rotation to some extent. Mars would be the same way.

ghwellsjr said:
Does the fact that the Moon, Mercury and Venus have synchronous rotations disqualify them from the same effect?

First, a note: Mercury actually doesn't keep one face to the Sun; it is in a slightly different tidally locked state, with 3 rotations (relative to the stars) for every 2 revolutions about the Sun. See the Wiki page:

http://en.wikipedia.org/wiki/Mercury_(planet [Broken])

That said, the effect I described is due to the rotation of an object relative to an inertial frame, not relative to any object it might be revolving around, so it would apply to the Moon and Venus (as well as Mercury, of course). Of course, the effect is smaller the slower the rotation period is.
 
Last edited by a moderator:
  • #13
LastOneStanding said:
Well, it's a bit unfair to fault Einstein for failing to take into account the effects of general relativity before he discovered it.:biggrin:
I agree with you but the first time I brought this up I got an argument:

https://www.physicsforums.com/showthread.php?t=558360
 
  • #14
PeterDonis said:
Yes, at least to a first approximation. As I said, the Earth is not exactly an oblate spheroid, because its substance, being mostly solid (part of the core is liquid), can resist the stresses set up by rotation to some extent. Mars would be the same way.

First, a note: Mercury actually doesn't keep one face to the Sun; it is in a slightly different tidally locked state, with 3 rotations (relative to the stars) for every 2 revolutions about the Sun. See the Wiki page:

http://en.wikipedia.org/wiki/Mercury_(planet [Broken])

That said, the effect I described is due to the rotation of an object relative to an inertial frame, not relative to any object it might be revolving around, so it would apply to the Moon and Venus (as well as Mercury, of course). Of course, the effect is smaller the slower the rotation period is.
Thanks. I would still call Mercury's rotation synchronous but I was mistaken about Venus'. It's very slow but not synchronized to the sun.

Your point about the equipotential surface of a planet causing clocks to run at the same rate was also made by PAllen in the thread I linked to in my previous post but I didn't catch the significance of it then. Anyway, thanks, guys.
 
Last edited by a moderator:
  • #15
ghwellsjr said:
I would still call Mercury's rotation synchronous

I don't think that's the standard usage of that term, but I'm not really familiar with the terminology in this area, so I don't know how the various possible tidally locked states are described. Somebody on the Astronomy forum probably does.

ghwellsjr said:
I was mistaken about Venus'. It's very slow but not synchronized to the sun.

Yes, I missed that too. Venus is even weirder in that its rotation is retrograde (and the period is slower than its orbital period about the Sun).
 

What is "Earth-Centered Time Reference"?

Earth-Centered Time Reference refers to the measurement of time based on the Earth's rotation and orbit around the sun. It is a system of timekeeping that is centered around the Earth, as opposed to other systems that may be based on the movements of other celestial bodies.

How is Earth-Centered Time Reference used in science?

Earth-Centered Time Reference is used in many scientific fields, such as astronomy, geology, and environmental science. It is used to track and measure the Earth's movements, as well as to understand and predict natural phenomena such as seasons, tides, and eclipses.

What are the advantages of using Earth-Centered Time Reference?

One of the main advantages of using Earth-Centered Time Reference is its consistency. The Earth's rotation and orbit are relatively stable and predictable, making it a reliable reference for measuring time. Additionally, it is a universal system that is used by people all over the world, making it a convenient and standardized way of keeping track of time.

How is Earth-Centered Time Reference different from other time systems?

Earth-Centered Time Reference differs from other time systems in that it is based on the Earth's movements rather than those of other celestial bodies. It also uses the concept of a solar day, which is defined as the time it takes for the Earth to make one full rotation on its axis, as its basic unit of time.

Are there any limitations or challenges associated with Earth-Centered Time Reference?

One limitation of Earth-Centered Time Reference is that it does not account for variations in the Earth's rotation, such as seasonal changes or long-term changes in the Earth's rotation speed. This can lead to slight discrepancies in time measurements over long periods. Additionally, Earth-Centered Time Reference may not be as precise as other time systems, such as atomic time, which is based on the vibrations of atoms.

Similar threads

  • Special and General Relativity
Replies
8
Views
1K
  • Special and General Relativity
3
Replies
70
Views
3K
  • Special and General Relativity
Replies
21
Views
1K
  • Special and General Relativity
Replies
10
Views
1K
  • Special and General Relativity
Replies
3
Views
1K
  • Special and General Relativity
Replies
8
Views
1K
  • Special and General Relativity
Replies
12
Views
1K
  • Special and General Relativity
Replies
14
Views
2K
  • Special and General Relativity
Replies
4
Views
1K
Back
Top