Hi Rob,
I really should have written "polar motion". This refers to the
way in which the earth's rotational axis moves around a little.
If you found the current axis at, say, the South Pole, dug down
under the ice, and put a marker in the crust, you'd find that
a bit later, your marker no longer coincided with the earth's
axis of rotation. This is described somewhat better at
http://en.wikipedia.org/wiki/Polar_motion
The basic idea is that if you observed your latitude and
longitude with sufficient accuracy, you would find that they are
not really entirely fixed quantities; they vary as the earth's
axis moves around.
However, this may not matter very much. It looks to me as
if the amount of the motion is so small that it should have no
serious effect at the level of radar measurements. (It would matter
a lot if I was trying to get earth-orbiting satellite positions down
to the last meter, for example; or if I was analyzing lunar
laser reflector data, which is good to a few centimeters.)
Jim Baer has sent me some papers, which I'm hoping will give
me a better handle as to which things matter and which don't.
As you mention, one does have to handle DUT1, which my code
does (I have a "td_minus_ut" function, and a "td_minus_utc"
function, with the difference meticulously observed throughout
the code.) One really has to consider this for fast-movers,
even with optical data, and you'd be in hard shape if you ignored
it and tried to handle radar data. I think the precession/nutation
routines I currently have will be accurate enough, but will have
to give them another look to be sure.
Quite aside from using the series solutions for precession/nutation,
I'll probably have to consider observed departures from the series
solutions. The IERS (International Earth Rotation Service) keeps
track of the earth's "real", observed rotation and distributes data
(which includes the amount of polar motion). I don't think it'll be
all that tricky to use.
-- Bill
Matson, Robert D. wrote:
> Hi Bill,
>
>> ... To do it _really_ right, I'd have to modify Find_Orb's concept
>> of the earth's orientation. It accounts for nutation and precession
>> and Delta-T, of course, but not for "polar wander".
>
> I'm not sure what additional effect you mean by polar wander.
> Precession and nutation should completely account for our pole's
> motion. (In reality, precession and nutation represent a somewhat
> arbitrary separation of very long period, and shorter period
> oscillations, respectively, in the earth's orientation. A single
> set of equations can cover all polar motions, but for historical
> reasons they tend to be treated as distinct from one another.)
>
> How many terms are you using in your nutation equations? Or more
> to the point, how small do the terms have to get before you no
> longer care about them in FindOrb? 100 mas? 10 mas? Smaller?
> Another factor you didn't mention is J2000 frame bias -- the
> small correction that must be applied to J2000 catalog coordinates
> to convert them to ICRS. Also, I assume you must be accounting for
> DUT1 somewhere in your code, which requires periodic updates. Not
> that important for MBAs, but it can affect astrometry for
> fast-moving, nearby objects.
>
> Best,
> Rob