# Time dilation for the Earth's orbit around the Sun

• I

## Main Question or Discussion Point

If we have 2 atomic clocks on Earth's orbit around the Sun, one on Earth's surface, at one pole, and the other on a spaceship, far from Earth, but traveling with the same speed around the Sun, the clocks would suffer the same kinematic time dilation or not?

I'm asking this because the clock on Earth would be stationary in the Earth's, co-moving, gravity well, while the spaceship clock would move (quite fast) through Sun's gravity well.

Last edited:

Related Special and General Relativity News on Phys.org
pervect
Staff Emeritus
To measure time dilation, one needs a way of comparing clocks at two different locations. In special and general relativity, this process is not universal. This is called "the relativity of simultaneity". To define the process of clock comparison, it is sufficient to define a specific coordinate system, with "simultaneous" events having the same value of time coordinate. The natural coordinate system to use in this circumstance would be some version of Schwarzschild coordinates, but other choices are possible.

In addition, this particular problem requires General relativity, and not just special relativity. General relativity includes gravitational time dilation, and not just kinematic time dilation. General relativity does have formulae that allows one to compute time dilation if one specifies what one means by time dilation sufficiently well (for instance by specifying some particular coordinate system as above). However, one does not get the correct result by using the SR formulae for time dilation in this circumstance.

Given a choice of coordinate system, time dilation can conveniently be defined as the ratio of proper time to coordinate time for an observer following some specific path (worldline).

Using the metric associated with the chosen coordinate system, one can compute the proper time from the relation ##d\tau^2 = g_{ij} dx^i dx^k##, where ##\tau## is the proper time, the ##x^i## are the choosen coordinates, and ##g_{ij}## are the metric coefficients associated with the particular coordinate choice.

This may seem slightly elaborate, but it's a safer way of doing the computating that makes it clear what assumptions are needed to answer the question. Sometimes assumptoins are not shared between the author of the question and the person who computes the answer, which can lead to confusion.

• DanMP and Dale
PeroK
Homework Helper
Gold Member
If we have 2 atomic clocks on Earth's orbit around the Sun, one on Earth's surface, at one pole, and the other on a spaceship, far from Earth, but traveling with the same speed around the Sun, the clocks would suffer the same kinematic time dilation or not?

I'm asking this because the clock on Earth would be stationary in the Earth's, co-moving, gravity well, while the spaceship clock would move (quite fast) through Sun's gravity well.
In your example, you could take a clock at a large distance from the Solar system, at rest with respect to the Sun as a reference clock. There are then three factors affecting the time dilation of a clock in the Solar system (if we ignore other bodies in the Solar system and consider only the Sun and the Earth):

The Sun's gravitational potential; the Earth's gravitational potential; the speed of the clock relative to the Sun (and the distant reference clock). Therefore, the time dilation for a clock on the Earth's surface should be greater than a clock somewhere else on the Earth's orbit, as it has an extra factor. But, the "kinematic" component of that time dilation in this coordinate system would be the same for both clocks.

this particular problem requires General relativity, and not just special relativity
I know that.

Given a choice of coordinate system, time dilation can conveniently be defined as the ratio of proper time to coordinate time for an observer following some specific path (worldline).

Using the metric associated with the chosen coordinate system, one can compute the proper time from the relation dτ2=gijdxidxkdτ2=gijdxidxkd\tau^2 = g_{ij} dx^i dx^k, where ττ\tau is the proper time, the xixix^i are the choosen coordinates, and gijgijg_{ij} are the metric coefficients associated with the particular coordinate choice.
This sounds very promising but unfortunately I don't have the skills to do the computations myself. I hoped that for the experts in this forum would be easy enough to do it and find the influence of the speed around the Sun on the clocks considered in the OP.

Sometimes assumptions are not shared between the author of the question and the person who computes the answer, which can lead to confusion.
What do you mean?

... the "kinematic" component of that time dilation in this coordinate system would be the same for both clocks.
I wouldn't be so sure, not without a proper computation, the kind that pervect suggested.

Dale
Mentor
I hoped that for the experts in this forum would be easy enough to do it
To do it quantitatively is not easy. Your question has no known analytical solution so it would need to be done numerically.

I wouldn't be so sure, not without a proper computation, the kind that pervect suggested.
This annoys me. You know that it is too difficult for you. You ignorantly assume that it should be easy for others because they know more than you. But when they (whom you recognize know more than you) tell you the qualitative result of the computation you petulantly reject your assessment of their expertise and demand a full computation.

I agree fully with his assessment of the outcome based on my previous experience working with such problems, but I also recognize that calculating the result quantitatively will not be easy.

You have no rational basis for rejecting the assessment. You are asking (demanding) a lot of work for no compensation. The result of the calculation is a difficult to obtain piece of trivia for anyone besides yourself. But you are making no attempt to do any of it on your own.

If you really want this done why don’t you either put in the work yourself to learn the material and do the computation or pay someone for the large amount of work required to do it.

Last edited:
To do it quantitatively is not easy. Your question has no known analytical solution so it would need to be done numerically.
If my question has no known analytical solution, maybe it would be a good exercise and even a material worth publishing ...

jbriggs444
Homework Helper
2019 Award
If my question has no known analytical solution, maybe it would be a good exercise and even a material worth publishing ...
The word "trivia" in @Dale's post is relevant here. Numerical solutions are like grunt work. Time consuming but not noteworthy.

Lots of problems have no known analytical solution. They are the rule rather than the exception. For instance, in classical mechanics, the three body problem.

You ignorantly assume that it should be easy for others because they know more than you.
Numerical solutions are like grunt work. Time consuming but not noteworthy.
I'm sorry for not knowing that it would be "grunt work" & "time consuming". I honestly believed that it should be fairly easy for experts.

I don't quite understand why numerically? And if it has to be numerical, why not using a computer to do it?

On the other hand, with a circular orbit and constant speed would be easier?

jbriggs444
Homework Helper
2019 Award
I don't quite understand why numerically? And if it has to be numerical, why not using a computer to do it?
Using a computer is what "numerical" means. More specifically, it tends to involve a differential equation solver using a method such as Runge Kutta.

PeroK
Homework Helper
Gold Member
I'm sorry for not knowing that it would be "grunt work" & "time consuming". I honestly believed that it should be fairly easy for experts.

I don't quite understand why numerically? And if it has to be numerical, why not using a computer to do it?

On the other hand, with a circular orbit and constant speed would be easier?
It's not too hard to see that for low speeds and weak gravity the time dilation is linear, in the sense that you can add "gravitational" and "kinematic" components. You can do this by considering four scenarios:

1) A clock at rest relative to the Sun.
2) A clock in the Earth's orbit relative to the Sun.
3) A clock on the Earth, with the Earth held at rest relative to the Sun.
4) A clock on the Earth, while the Earth orbits the Sun.

By considering these in relation to each other and in relation to a reference clock at rest and a long way from the Sun, you can see how to add the "components" of time dilation in each case. This was my argument anyway.

However, the Einstein Field Equations are non-linear, hence the solutions for two massive bodies don't, in general, superpose linearly. So, there is at least a small correction to the simple analysis above; even for low speed solar orbits.

Crunching that is not particularly interesting because a) we have a good approximation already; b) it doesn't tell us anything significant about GR; and, c) in general things like "kinematic" and "gravitational" time dilation are artefacts of the coordinates and have no physical significance in any case!

Dale
Mentor
I don't quite understand why numerically?
There is not an analytical solution for the spacetime of a massive body orbiting another massive body. And unfortunately the EFE is nonlinear so you cannot simply add the two single body analytical solutions to get a valid overall solution. So it requires a numerical solution or some simplifying approximations.

t's not too hard to see that for low speeds and weak gravity the time dilation is linear, in the sense that you can add "gravitational" and "kinematic" components.
You mean something like this?

However, the Einstein Field Equations are non-linear, hence the solutions for two massive bodies don't, in general, superpose linearly. So, there is at least a small correction to the simple analysis above; even for low speed solar orbits.
You said/implied that "at least a small correction" has to be expected. What if the "correction" is not small?

Crunching that is not particularly interesting because a) we have a good approximation already; b) it doesn't tell us anything significant about GR; and, c) in general things like "kinematic" and "gravitational" time dilation are artefacts of the coordinates and have no physical significance in any case!
You (all) are not interested to see how being static in a co-moving "gravity well" may affect the "kinematic component" of time dilation?

There is not an analytical solution for the spacetime of a massive body orbiting another massive body. And unfortunately the EFE is nonlinear so you cannot simply add the two single body analytical solutions to get a valid overall solution. So it requires a numerical solution or some simplifying approximations.
I suggested to consider the orbit a circle. If you know more "simplifying approximations", please post them.

Dale
Mentor
I suggested to consider the orbit a circle. If you know more "simplifying approximations", please post them.
That wouldn’t simplify much. The big simplification would be to use the weak field approximation, or equivalently to use a linear approximation to the EFE.

The big simplification would be to use the weak field approximation, or equivalently to use a linear approximation to the EFE.
I don't know if this would be appropriate.

Using a computer is what "numerical" means.
So why/how is this a "grunt work" if the computer does it?

Dale
Mentor
I don't know if this would be appropriate.
Hence the need for a lot of work

So why/how is this a "grunt work" if the computer does it?
The computer won’t program itself

• PeroK and jbriggs444
Nugatory
Mentor
So why/how is this a "grunt work" if the computer does it?
It's a lot of work to get the differential equations and initial conditions into a form suitable for the computer; and then when the computer comes back with a number, you have only the same understanding of the physics involved that you started with.

• jbriggs444
The computer won’t program itself
I thought that there are already made programs for EFE and/or problems like this.

If you really want this done why don’t you either put in the work yourself to learn the material and do the computation or pay someone for the large amount of work required to do it.
How much would it cost?

Nugatory
Mentor
You said/implied that "at least a small correction" has to be expected. What if the "correction" is not small?
If the correction is not small, then the method @PeroK outlined isn't usable for this problem.

A more mundane example: We have two points on the surface of the earth, one five meters due east of a stake in the ground and the other five meters due west of the stake. You ask me to calculate the straight-line distance between them; I say that it's easy because distances just add, so it's ten meters. But in fact there is a small correction required because of the earth's curvature - the two five-meter segments are actually two sides of a triangle and we're looking for the length of the third side. So is my add-the-lengths method an acceptable way of solving the problem? It depends on whether we care about the small correction - but it's there either way.
You (all) are not interested to see how being static in a co-moving "gravity well" may affect the "kinematic component" of time dilation?
That question is less interesting than it first appears, because"static" and "co-moving" are just (again, @PeroK) artifacts of the coordinate system.

Suppose that the clock in the gravity well is emitting a flash of light once every second so that we can watch it ticking from a distance by observing the arrival of the flashes. We will not, in general, receive the flashes at a rate of one per second; this will be a combination of time dilation and Doppler effect, and we can further divide the time dilation into a kinematic and a gravitational component. However, I can make this split come out pretty much any way that I please by choosing coordinates appropriately; all the real physics is in the ratio between the proper time elapsed between two emission events and the proper time elapsed between the two corresponding detection events.

• PeroK
Dale
Mentor
I thought that there are already made programs for EFE and/or problems like this.
Not that I am aware of. Most labs doing this kind of research write their own code.

How much would it cost?
I would charge $60/hr. That is less than my time is worth, but I am not highly qualified so I cannot expect to charge full price. this will be a combination of time dilation and Doppler effect, and we can further divide the time dilation into a kinematic and a gravitational component. However, I can make this split come out pretty much any way that I please by choosing coordinates appropriately; all the real physics is in the ratio of proper time elapsed between two emission events and the proper time elapsed between the two corresponding detection events. Consider the problem in another way: the spaceship clock starts from the Earth, where is synchronized with the other one, then flies off the Earth, going on orbit around the Sun, with the same speed as the Earth, but far from it (few millions km away). After a while (months, years), the spaceship returns and the reunited clocks are compared. When I wrote "kinematic component" of time dilation I meant the influence of the orbital speed around the Sun in the calculation of the elapsed time between start and finish, for each clock. I would charge$60/hr. That is less than my time is worth, but I am not highly qualified so I cannot expect to charge full price.
So I may expect more. Anyway, the next/big question is how many hours an expert would need to complete the task?

Nugatory
Mentor
When I wrote "kinematic component" of time dilation I meant the influence of the orbital speed around the Sun in the calculation of the elapsed time between start and finish, for each clock.
The difference in elapsed time is the difference in the length of the world lines of the two clocks along their respective paths between the separation event and the reunion event. How much of this we attribute to the orbital speed and how much we attribute to gravitational effects is arbitrary.

Ibix
How much of this we attribute to the orbital speed and how much we attribute to gravitational effects is arbitrary.
Indeed. Isn't the point here that we either accept @PeroK's linear approximation, in which case the answer is trivial, or we don't, in which case the spacetime isn't static and the answer is a matter of personal preference?

• DanMP
pervect
Staff Emeritus
Just for clarity, I'd advocate using the same linear approach that has been suggested. Doing a full-on GR computation just isn't sensible, the linear approximation is good enough and is widely used in any and all solar system problems.

The tricky part is that the problem is specified in a manner that makes it coordinate dependent, so one needs to choose a coordinate system. There are two main general philosophies here - geocentric and barycentric, i.e. earth-centered and sun-centered (actually the center of mass of the solar system, slightly different from sun-centered).

But it's unclear specifically which coordinate system to to recommend. PPN because it's in a lot of old textbooks? ICRS (or, for the Earth centered case, IERS) as being modern? What about BCRS and GCRS - the former of which I happen to have a line element for which would expedite the computation.

It would actually be better if the problem were specified in a manner that wasn't coordinate dependent, but the OP seems narrowly focused on doing things "their way" rather than learning new things.

I wouldn't actually expect much difference between any of the various coordinate choices by the way.

I''m also not feeling motivated to work through all the details - I don't actually think they'd help the OP, who in my opinion would be better off widening their focus. I am willing to comment a bit on the general approach that I'd use, which I've already done.

• PeroK
... we either accept @PeroK's linear approximation, in which case the answer is trivial, or we don't, in which case the spacetime isn't static and the answer is a matter of personal preference?
By "spacetime isn't static" you mean linear frame dragging? It is possible? And it means that the clock on the Earth would not be affected at all by the orbital speed (of the Earth) around the Sun?

It would actually be better if the problem were specified in a manner that wasn't coordinate dependent, but the OP seems narrowly focused on doing things "their way" rather than learning new things.
In a way it's true, I'm focused on learning if GR would predict/allow linear frame dragging.