Two years ago this month I was driving E of Greensboro NC on I-85,
before dusk, and the moon had just risen in the east (summer 2008).
It looked as big as any I have seen on the horizon, and since I had a
digital camera with me, decided to take some pictures. I ran across
them tonight, and decided to finally look at them in detail.
On the horizon, I took 1 picture with the optical zoom all the way
out, and 1 picture with the the optical zoom all the way in. 7 hrs
later that night, at my house perhaps 30-40 miles east of the earlier
pictures, I again used the same camera to take 2 pictures at zenith,
again 1 with the optical zoom all the way out, and 1 with the the
optical zoom all the way in.
Based on geometry, and ignoring refractive effects that influence the
perceived size, the angular extent of the moon at zenith is somewhere
in the neighborhood of 1.5-2% larger than at horizon, and expected
that to be reflected in the pictures, thus lending credence to the
accepted notion that the large "horizon moon" is an illusion. I was
surprised instead to find a significant discrepancy in the pictures.
They confirmed that the moon was larger in extent at zenith, but the
magnitude of the difference was quite significant.
Zoomed in, the zenith picture showed the moon to be easily a factor
of 2 or even slightly more than the extent of the horizon moon.
Zoomed out, the zenith picture showed the moon to be about 20-30%
larger than the horizon moon.
I chose the zoomed-in and zoomed-out camera settings because they
were repeatable. Also, the pictures all have the same number of
pixels in them. This in principle should normalize my ability to
merely compare the moon's disk from picture to picture. I'm trying to
figure out what effects might be causing the unexpected result.
One possibility is that I simply messed up, and I took 4 pictures at
unknown zooms. I seriously seriously doubt that, but can't prove it.
Another is that the zenith pictures, taken in darkness and with no
special precautions as to exposure, exhibit blurring and washout due
to misfocus etc. However, the night/zenith shots look pretty darn
good - the moon's disk is quite round, and the focus actually doesn't
look bad.
More sophisticated is the theory that due to overexposure, misfocus,
and instability, the cross-sectional intensities of the night-time
zenith pictures are bell-curve-like, and my eyes can't appropriate
judge the diameter, requiring a look at the bell-curve itself to
determine an effective angular extent. Fine, I have to try looking at
the cross-sectional intensities to make sure, but my intuition is
telling me this is not the issue.
Wondering if anyone had actually *done* this for themselves, and had
observations. I'm working on pasting the images side by side to be
looked at if anyone wants to see them.