On 2016-10-07 16:48,
norm_hecht@... [find_orb] wrote:
>
> I'd like to determine the orbit of Earth's moon. Is Find_Orb a good
> tool for this project? I'll actually be doing observations via an
> iPhone; the camera is good enough to observe the moon (at least as
> a blob on screen), and the device can capture its orientation, GPS
> location, and the time of the observation.
Hi Norm,
I rather like this idea. I could imagine people getting phone camera
"astrometry" of one sort or another for the Moon, and/or perhaps Venus
(dunno just how faint various breeds of smartphone will go), and computing
an orbit from the observations. The idea of doing it with the alt/az data
measured from a smartphone is appealing, though I'd wonder about the
accuracy of that data; if it's good to within a degree, I'd be rather
impressed.
But even if that method didn't work, I could imagine some reasonably
simple methods for measuring lunar or planetary positions well enough
to get good orbits (*). It strikes me as an excellent educational project.
You should, without too much trouble, be able to get the length of
the lunar month, and the fact that the lunar orbit is elliptical,
and so forth.
I thought I'd have to make some small modifications to Find_Orb before
it could do something like this. However, I fed it some totally artificial
astrometry "taken" by Peter Birtwhistle at (J95) Great Shefford and by
(703) Catalina Sky Survey (actually computed by asking Guide for lunar
RA/decs at various semi-random times with the Moon up and the Sun down),
and got the orbit right away :
http://www.projectpluto.com/temp/mpec.htm
(If you load the astrometry at the above URL into Find_Orb, it will
first decide that this is an object orbiting the Moon a few hundred
meters from its center. You'll have to go into the Settings dialog and
tell Find_Orb that you want a geocentric orbit. Do that, and you should
get the result shown in the pseudo-MPEC.)
> Right now I'm using some functions from Find_Orb to convert azimuth
> and elevation angles to right ascension and declination, although
> it looks like I'm not using them correctly: October 7, 2016,
> 20:14 UTC become Julian day 2457669.36330594, which day_to_dmy
> turns into 2016, 9, 24.
Hmmm... 13 days is the difference (during our lifetimes) between
the Julian and Gregorian calendars. That JD does match 2016 Oct 7
(Gregorian) and 2016 Sep 24 (Julian). When you call day_to_dmy,
the last argument should be either CALENDAR_GREGORIAN (if you want
to use Gregorian for all dates) or CALENDAR_JULIAN_GREGORIAN (if
you want 1582 October 4 to be followed by 1582 October 15).
-- Bill
(*) The first and simplest measurement method I thought of was rise/set
times. In theory, time the rising and/or setting of an object six
times, and you've got the data needed to compute its orbit. The
accuracy will be better if those times are spread out a little bit,
and it would probably help to measure them from different points
on the earth (should be easy enough to collaborate with others on
this).
The big drawback I see with this is that rise/set times vary a bit
due to atmospheric conditions, and a difference of one minute would
correspond to a quarter of a degree. Probably a lot better than a
smartphone camera could do, but still not great.
Second method would be to measure the apparent angular distance
between the limb of the moon and a bright star: "The Moon's limb
was 6.3 degrees from Regulus at thus-and-such time/date from this
place." Again, six observations would get you an orbit. This
might work somewhat better; it shouldn't be too hard to build
a pseudo-sextant to get that level of accuracy.
Either of these would require some modifications to Find_Orb
to allow "observations" to be, not just RA/decs, but alt/azzes
or angular distances from a known RA/dec.
Third method would be to just take pictures with a smartphone
and run them through astrometry.net to get for-real RA/decs.