Error trying to compute altaz from earth position to jupiter

[sizzzzlerz]

My ultimate goal is to compute the angles between the largest moons of Jupiter and Jupiter. To facilitate that, I loaded the jup365.bsp Spice kernel which includes Jupiter, the sun and Earth and a number of moons. To familiarize myself with this file, I first tried computing the alt and az from my location (home) to jupiter using home.at(t).observe(jupiter).apparent().altaz(). What I got was a KeyError exception that basically stated that Saturn was not found in the kernel. Why does Saturn need to be in the kernel? Is this a possible bug?

[Bernmeister]

Although I cannot offer specific help, I'm confident the first thing those that can help you will ask for a small complete test script which demonstrates the error.

[sizzzzlerz]

Here is a script that demonstrates what I'm seeing. Note that the jup365.bsp is over 1GB so if you don't already have it, it is going to take some time to download.

from skyfield.api import load, wgs84

ts = load.timescale()
t = ts.now()

# change the path here as appropriate
eph = load("/Users/lskea/Downloads/jup365.bsp")
print(eph)

earth = eph['earth']
jupiter = eph['jupiter']

astrometric = earth.at(t).observe(jupiter)

ra,dec,dist = astrometric.radec()
print(ra, dec, dist)

altaz = astrometric.apparent().altaz()

[labonocl]

Hei,

when calling apparent() on a position, according to the docs the following happens:

This applies two effects to the position that arise from
relativity and shift slightly where the other body will appear
in the sky: the deflection that the image will experience if its
light passes close to large masses in the Solar System, and the
aberration of light caused by the observer's own velocity.

And indeed, the code always applies the deflection of light due to the Sun, Jupiter, and Saturn. In relativity.py it defines deflectors

deflectors = ['sun', 'jupiter', 'saturn', 'moon', 'venus', 'uranus', 'neptune']

And the add_deflection() function actually has a keyword called count, that is set to 3 as default, i.e., the first three elements (Sun, Jupiter, Saturn) are included. The function apparent() currently has no way of controlling the count keyword.

So I see a few solutions:

• Load Saturn's ephemeris also
• change your local skyfield code to either set the count default value to 2 (in add_deflection() in relativity.py), or add a count keyword to apparent() (in positionlib.py) which passes it on to add_deflection()

From a skyfield development point-of-view, issuing a warning when a deflection cannot be applied due to lacking ephemeris is probably the cleanest way.

BTW: as it is now, your code will not work, as altaz() needs to be called from a position on Earth (i.e. a latitude/longitude). But that is easily fixed. Also, it is the apparent() call that causes the error, not the altaz() call.

[labonocl]

Regarding my first point, the following works:

from skyfield.api import load, wgs84, N, W

ts = load.timescale()
t = ts.now()

# change the path here as appropriate
eph = load("jup365.bsp")
bodies = load('de421.bsp')
# push Saturn's ephemeris in there - not sure how safe this is
eph.spk.segments.append(bodies.spk.segments[5])
eph.segments.append(bodies.segments[5])
eph.codes.add(6)

print(eph)

earth = eph['earth']
jupiter = eph['jupiter']

boston = earth + wgs84.latlon(42.3583 * N, 71.0603 * W, elevation_m=43)
astrometric = boston.at(t).observe(jupiter)

ra,dec,dist = astrometric.radec()
print(ra, dec, dist)

altaz = astrometric.apparent().altaz()
print(altaz)

BTW, I also calculated the position without Saturn's contribution to the deflection, and (as expected?) in this case the difference in angle is of the order of 1e-8 degrees.

[sizzzzlerz]

I caught the mistake about calling altaz on apparent after I'd hurriedly typed in and submitted here what I'd been running in iPython which did correctly make the call to apparent. I just missed that line when searching for the lines to enter into the script. Thanks for catching here however.

Maybe you or someone else can address why perturbations from Saturn would affect light traveling to Jupiter. Won't Jupiter always be in front of Saturn or am I missing some scenario where it would have an affect?

[labonocl]

The image above sketches a kind of worst case scenario. The orbits of the three planets Earth (E), Jupiter (J), Saturn (S) are to scale, the planet sizes obvisouly not. Observing Jupiter from Earth (without Saturn), the light would travel the dotted line. However, because of Saturn's mass, the light travels along a curved path (highly exaggerated in this schematic), such that you would assume Jupiter is where the dashed line points.
However, two comments:

1. According to your original post, you are interested in computing "the angles between the largest moons of Jupiter and Jupiter". Since the moons are very close to Jupiter (in comparisson to the distance to Saturn), the light paths from the moons and Jupiter itself will be affected by Saturn in essentially the same way, such that the angles between moons and Jupiter will likely not be affected (much).
2. Obviously it all depends on your requirements, but putting in some numbers for the expected deflection in the case sketched above, I get the deflection angle to be about 1 micro arcsecond, which is (well) below the angular resolution of current instruments. But again, it all depends on the accuracy that you require for your application.

[sizzzzlerz]

Thanks for your explanation. My requirements are I just want to draw pretty figures of the movement of these moons around Jupiter like you find in various astronomy magazines. I really don't think a 1 microsecond discrepancy is going to significantly alter them 😁