A blog about a telescopic system at the Mauna Loa Observatory on Hawaii to determine terrestrial albedo by earthshine observations. Feasible thanks to sheer determination.
9 Degrees C. Some thin cloud from the start – then more and more.
Lots of Moon observations in ‘dither’ mode, through all filters.
Day started with CCD flatlining. Reboot. Then all OK, Still have to set sub-degree offsets on pointing to put Moon in middle. Why? Will test the MOVETOREF command.
FW apparently OK – no noticeable ‘sticks’.
Peter sent Chris a file called “ROLO_765nm_Vega_psf.dat”
which contains the PSF of Vega measured by Tom Stone.
Here is the profile:
the horizontal scale is log of r in arsecs.
the inner part is something like a Gaussian, like King finds.
the halo outer part goes at r^-2.4, in the range 1.2 < log(r) < 2.0,
so steeper than King but not as steep as our steepest profiles
to date (these are r^-3).
The very outer parts are certainly affected by sky subtraction,
as the light will be well below sky at log(r)>2, one could fix this
probably by choosing the sky so that the halo continues at the
same gradient, but this is of course arbitrary.
It seems that for a wide range of telescopes and instruments,
the distant stellar light falls off as a power law — we see this
in the ES telescope, a 35 mm camera, King’s profile, and a
range of other instruments/telescopes summarised in a paper
by Bernstein (ApJ, 666, 663, 2007).
The slope of the power law can change even on the same
telescope/instrument! We see different powers on different
nights. This is important!
Attached plot is the Altair profile from the latest run.
The data don’t go quite as deep as Menkab, but are highly
interesting all the same.
for the IR, V and VE2 filters, the slope is very close
to our canonical r^-3 or so which we get for the moon.
it was a very clear night, so that’s SUPER!
for the B and VE1 filters the halo is rather different — as
can be see in the plot!
we might have some sky subtraction problem with VE2, so let’s
not worry about that too much at this point.
the peculiar profile is definitely B — as before with Menkab —
there is a peculiar hump in the psf at log(r) ~ 0.8 to 1.0
(6 to 10 pixels). We saw the same in Menkab and wondered
about focus.
Could there be something funny going in with internal
reflections in the B filter, so that an extra halo appears around
stars in that filter but not in the others — or at least not
prominently?
We obtained images of the full moon from Sydney with the Canon
35 mm camera, reaching out to about 15 degrees from the moon
— three times further out than the King profile reaches. We are
curious about how far out it goes as a power law.
The plot shows the profile done for a range of exposure
times, getting longer and longer, at the cost of overexposing the
moon and halo around it, but better sampling the faint halo
far away.
I have shown the raw data merging into the background sky in the
bottom panels. Subtracting this sky gives the halo only profile in
the middle panel for the range of exposures.
Top panel shows the overall profile by using only the valid (not
overexposed) parts of each curve — by simply accounting for the
different exposure times — nothing more has been done. They
sit on a pretty well defined locus all the way from the edge of the
moon out to the edge of the frame (~15 degrees out) and follow
roughly r^-1.8. It’s probably flatter than this closer in. This is of
course the scattered light, not the PSF per se, but very far from the
moon — like > a few degrees — the moon is essentially a point source
anyway (it’s only 30 pixels across on a chip 3900 pixels wide).
I learned something important! The profile is r^-1.8 when sky subtracted!
We can follow it well beyond the point that the halo is much fainter than
the sky. So King I think is talking about the sky subtracted profile as well.
I have been wondering when the King profile would actually join
the sky and become a DC component… of course it doesn’t, that’s
the point. The DC component, i.e. Rayleigh scattering (?) is caused
by the bright source as well, but is not considered part of the profile.
I guess. So if we were trying to model the scattering over very large angles
we’d have to consider both components. My guess is that for our
ES telescope the field of view is so small that the scattered light is
the halo only — the Rayleigh component is negligible. One starts
to see it in the 35mm camera data at about 1000 pixels (1 pix
is approx 1 minute in these data) so about 10 degrees to 15
degrees from the moon (lower panel shows a blue curve which
essentially joins the bias level at large r, whereas the green and
black curves are flattening out to a level which is bias + a fair
bit of sky). In these data, the sky is about 5*2.5 = 7.5 magnitudes
below the full moon surface brightness (seen in upper panel, with
brightness of the moon at log(r)=0 and the (subtracted) sky having
similar brightness as the halo at about log(r)=3 — beyond which one
traces the halo below the sky).
